# Welcome to HEARTCOUNT HEARTCOUNT Public is a free, cloud-based data analytics tool that enables you to perform rapid, easy, code-free and in-depth exploratory data analysis(EDA) on your own data. {% hint style="info" %} To get free access to HEARTCOUNT Public, [Click Here](https://www.heartcount.io/en?utm_source=gitbook\&utm_medium=manual\&utm_campaign=main) {% endhint %} ## ## Explore Millions of Data Points at the Speed of Thought The core design principle of HEARTCOUNT is to let the users describe, explore, and summarize the given dataset at the speed of thought, not at the speed constrained by data architecture of the tool. Therefore, the user can completely forget about the underlying technology and mechanics, and solely focus on asking relevant questions to the data and interpreting what’s revealed in front of their eyes. {% embed url="" %} ### **- How Does It Work** HEARTCOUNT prepares and stores the complete dataset and and associated meta data in the browser’s memory so that the user can keep asking new questions(i.e., mapping new variables to the visual elements; sales variable to Y-axis, product\_category variable to X-axis, and other variables to color, size, etc.) without any noticeable delay. To further elaborate, our proprietary visualization engine interprets user’s question(i.e., data\_values to visual\_elements mapping) and renders the corresponding and context-aware visualization on the fly: **HEARTCOUNT is the only tool in the world that can interactively visualize over 10,000,000 individual records (instead of showing only the aggregated or randomly clipped data).** ### - See Both the Forest and Trees Understand the big picture then zoom into the relationship you want to dig deeper. You don't have to limit yourself to an aggregated representation of your data. As you freely explore the data at individual record level, you will encounter uncertainty inherent in your data and discover more pragmatic and non-obvious patterns. {% embed url="" %} ## What You Can Do with HEARTCOUNT HEARTCOUNT provides you a variety of easy and powerful data visualization technologies. Take a quick look at what HEARTCOUNT can do for you in this section. ### - Just Ask Your Data ![](https://1130022100-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FHmGL0fwQrDEm9mSq6FQN%2Fuploads%2Fk07nZwpjQT6J8ukY2Bjg%2Fmain_justask_v1%20$1$.gif?alt=media\&token=bd4fe30e-73d4-49f4-a1d8-540c77a494d7) You don't need to look far to determine which of HEARTCOUNT's potent features is ideal for your analytical queries. By using the "Ask Data" feature, HEARTCOUNT will instantly present the best visualization to address your questions. {% hint style="info" %} If you're interested in checking out the full features of premium HEARTCOUNT, please contact and get a free trial. {% endhint %} # HEARTCOUNT's data community, DDMA DDMA, short for DON'T DATA ME ALONE 😝, aims to dispel the existential loneliness often accompanying solitary data analysis. The **"**[**DON'T DATA ME ALONE (a.k.a DDMA)**](https://community.heartcount.io/en/)**"** community is a vibrant space designed to foster collaboration among data enthusiasts. Hosted under the umbrella of HEARTCOUNT, this community is geared towards unlocking the potential of data through collective efforts. The phrase "DON'T DATA ME ALONE" is a playful spin on words, here, 'Data' is utilized as a verb, leaning more towards being commanded to analyze data rather than actively choosing to do so. This phrase encapsulates the existential loneliness professionals often feel when left alone to analyze data. It's inspired by a comic titled "Don't Leave Me Alone," reflecting a narrative where a young girl from an immigrant family, facing unjust discrimination, yearns not to be left alone in a harsh reality, mirroring the longing for community and collaboration in the data analysis realm. ![](https://community.heartcount.io/en/content/images/2023/10/-----------2023-09-05-------10.35.43.png) I would suggest diving into the specific sections, webinars, and blog posts available on the community platform to provide readers with a comprehensive understanding of the resources and knowledge-sharing occurring in this thriving community. {% hint style="info" %} Join our brand new data community: [DON'T DATA ME ALONE](https://community.heartcount.io/en/) ! {% endhint %} # How to start Are you curious about what sets HEARTCOUNT Premium apart from the free version? If so, you can find out all the details on this page. ## HEARTCOUNT Premium enables you to save campaigns for ongoing analysis. * Easily **save your work with a click and manage campaigns at your convenience.** The number of saves you get depends on your plan, ranging from 15 to unlimited with Premium Plus. * For enhanced security, enable data obfuscation during campaign creation. * Bookmark valuable analysis results within campaigns for easy retrieval, even export visualizations as HTML files. ## Not just visualizations, but also get automated analytic functions. * **Register KPIs for unique automatic analysis and pattern discovery,** with the capacity to manage up to 5 or 10 KPIs based on your account type. * Understand KPI relationships with other variables in simple terms, strategize on maximizing/minimizing KPIs, and define distinct groups to compare data. ## Discover more with HEARTCOUNT Premium Start your 1-month free trial and unlock more value from your data with HEARTCOUNT Premium. Upgrade after experiencing the full suite of tools. Note that the free trial is limited to one per account. # Campaigns In order to use HEARTCOUNT's Advanced Analytics Service, you have to upload your "Campaign" first. 😊 ## 1. **Saving Campaigns** You can save campaigns when you have at least one campaign credit (while using a **Premium** Plan of HEARTCOUNT). The number of credits you have determines the number of campaigns you can save (for the **Premium Plus+** plan, there's no limit on the numbers of storable campaigns). Click on the blue '`Upload this Campaign`' button below to save your campaign on the HEARTCOUNT server. If you don't save your campaign, all the analysis will disappear once you close your browser.

Saved campaigns can be viewed and deleted from the 'Campaign Management' tab, and you can also switch between different campaigns. Each account can store up to 15 campaigns.

## **2. Advanced Settings** When saving a campaign, you can set the following advanced options: * **Data Obfuscation**: Choose whether to obfuscate (anonymize) the dataset before uploading it to the server. > **What is data obfuscation?** > > This feature automatically obfuscates sensitive data before it's uploaded to the cloud server, ensuring only the obfuscated data is stored. This allows users to analyze data with HEARTCOUNT confidently from anywhere. * **Upload Filtered Dataset**: If filtering is applied, you can choose to upload only the filtered records/variables or remove all filters and then upload. * **Recreate Derived Variables**: If the above option is ON, you can choose whether to recreate derived variables like bin/percentile, which are automatically generated by HEARTCOUNT, based on the filtered dataset.

## **3. Campaign Management** Think of campaign management as a folder on your computer used for organizing files, but in HEARTCOUNT, these files are campaigns. Register multiple files as campaigns, and then, whenever you wish, move between different campaigns, delete them, or organize the ones you've registered. This is the ideal tab for such tasks.

You can view the campaigns you've registered through the second tab on the left, **'Manage Campaign'**. 1. **Delete Campaign**: If too many campaigns are registered, it might overload the server. It's advisable to delete unnecessary campaigns. HEARTCOUNT recommends having no more than 15 campaigns per account. 2. **Activate Campaign**: When you select a campaign you want to analyze, you'll notice the active campaign name next to the user ID on the right changes. After that, you can start your desired analysis from the left menu.
# Managing KPIs ## **KPI Registration & Management** In HEARTCOUNT, think of a KPI as "the main business metric variable in our current analysis." For instance, if a campaign's goal is "finding rules for optimizing sales and profit," then the KPIs for that campaign would be sales and profit. While you can analyze most metrics in HEARTCOUNT even if they aren't set as KPIs, * SMART DISCOVERY Dashboard feature, which automatically identifies patterns in individual KPIs and recommends significant analysis results, is only available for metrics set as KPIs. 👉 Learn more about [SMART DISCOVERY](https://public.heartcount.io/advanced-analysis/dashboard) * Driver analysis and the Explainer function analyze the relationship between numerical KPIs and other variables. Therefore, a numerical KPI must be registered to utilize these features. {% hint style="info" %} HEARTCOUNT recommends registering no more than 10 KPIs per campaign. {% endhint %} You can view all columns (rows) of the previously uploaded data. From these columns, you can set the ones you want to focus on as KPIs and manually set the variables.

## **1. KPI Registration** ### **(1) Numerical Variables** Let's first look at how to set a numerical variable as a KPI. > * Numerical variables refer to data made up of continuous numbers. ex) Age, height, income, etc. > * Categorical variables refer to non-continuous data that can be divided into several categories or ranges of similar characteristics. ex) Gender, blood type, etc.

Numerical variables are automatically marked with a number (#) in the system. Choose one from the \[KPI Types] (score, money, time, others). Then, select one or both of the calculation methods: average or total. Note that the \[KPI Type] is only a classification standard and does not affect the analysis results. **1-1. Setting Derived Variables** As mentioned in campaign creation, HEARTCOUNT offers automatic features like categorizing numerical variables and **binning** them into ranges, as well as processing them into **percentiles**. Thus, you can create two derived variables. 👉 Learn more about the binning feature (link)

You can set up to 4 of these derived variables as KPIs. ### **(2) Categorical Variables** Categorical variables are those defined by non-numerical values for creating KPIs (YES/NO, Male/Female, etc.). You can select **up to 4 KPIs** per variable.

Once you've set up your KPIs, it's time to start your analysis in earnest. * By clicking on the dashboard icon right below the KPIs management icon on the left menu, you can directly view and proceed to the analysis function for the KPIs you've set. \ 👉 Go to the [Dashboard manual](https://public.heartcount.io/advanced-analysis/dashboard) * You can freely add, delete, or modify KPIs during the analysis. # Data Obfuscation {% hint style="info" %} Please note! \ Important Notice: The data obfuscation feature is exclusive to the **Premium (paid) version of HEARTCOUNT**. In the free version, datasets uploaded are not stored on our servers and will be automatically removed upon closing the browser. {% endhint %} ## **1. What is Data Obfuscation?** Many hesitate to adopt cloud services due to concerns about leaks of *sensitive/personal information (security issues)*. HEARTCOUNT offers a data obfuscation feature to ensure users can confidently analyze data online from anywhere. ## **2. Uploading Data to the Server after Obfuscation** When saving a campaign, you can click on advanced settings to turn the obfuscation feature on/off. If set to obfuscation, only the obfuscated file, not the original one, is stored on the server.

Let's try saving the campaign after checking the obfuscation option. * A message will pop up suggesting you download the modified data, which has been obfuscated, not the original data. Please click on "`download`".

* Upon successful execution, click "OK".

Now, instead of uploading your original data that might raise security concerns, you can upload the obfuscated data and utilize HeartCount. You have the option to decide whether to store the obfuscated data in the current browser memory. Curious about HeartCount's unique data obfuscation feature? (Click) # Smart Plot ## :question:What It Can Do for You Smart Plot is the key feature of HEARTCOUNT where typical code-heavy visualization tasks can be executed easily and come in handy. To visualize data in a way you want to examine it, all you have to do is to simply select which variable goes x-axis and which goes to y-axis and choose the type of visualizations you wish to use.

Smart Plot provides a variety of visualizations of your data, which include such below.

Data Type	Available Visualization
Between numeric and other numeric	Scatterplot
	Trend Line(regression line)
	Heat Scatter
Between categorical and numeric	Bar (average or sum)
	Stacked Bar
	Stacked Area
	95% confidence interval
	Boxplot
Between categorical and categorical	Ratio Bar Chart
	Stacked Count Bar
Between time series and numeric	Time Series Line Chart
	Stacked Area
	Trend Line
	Forecast

## :page\_with\_curl:How to Use ### - Basics Smart Plot consists of four key sections. You can easily create a suitable visualization of your dataset using these sections without writing a single line of code.

Area	What's It For
1. Main Area	This is where a plot will be displayed. You will be able to interact with plot elements such as data points in a scatterplot to further investigate the dataset to find an answer to your analytic inquiries.
2. Side Menu	This is where you may configure Smart Plot's parameters, such as which variables to use to change the colors or sizes of data points, or to filter the data. Also, you could choose
3. Variable Selection	This is where you can choose which variable to create a plot suitable for your analytic purpose. As with creating a data visualization in a code-heavy setting(R/Python), you must choose which variables you would place in the x and y axes and which to use for subgrouping or faceting.
4. Visualization Type	This is where you can choose which type of visualization you would use to plot the data. Given the variables for the x and y axes, the Visualization Type tab will provide you several options you could choose from to correctly visualize your data.

### - Types of Visualizations This section will discuss the many sorts of visualizations possible in relation to the specified x and y axes variables. {% tabs %} {% tab title="Two Numeric Variables" %} When you put numeric variables on both axes, a simple scatterplot will be displayed on Smart Plot. Also, a Pearson correlation coefficient will be given as a basic information on these two variables.

See in Detail

* There are two available types of additional visualizations for the scatterplot. * You may choose to display a **trend line**

, which is basically a regression line. It displays how the x and y variables are linearly correlated.

* On the left of the trend line icon lies a number. It is a Pearson correlation coefficient that shows how much those variables are linearly correlated. * Pearson correlation coefficients are on or between −1 and +1. * If it equals to zero, it means they are not linearly correlated at all. * If it equals to plus one, it means they are strongly positively correlated. * If it equals to minus one, it means they are strongly negatively correlated. * If you drag to select some data points of interest, it will show you the linear relationship of only **the selected data**.

* The other available option is **heatmap**

(often known as heatscatter). * This visualizes around which area data points cluster together the most.

* The gradient color scale will change in accordance with the color settings on right top of HEARTCOUNT.\

{% endtab %} {% tab title="Category on X & Numeric on Y" %} There are two major ways to visualize the data when you put a categorical variable on x-axis and a numeric variable on y-axis. One is to visualize individual data points and the other is to create subgroups based on the categorical variable on the x-axis.

Individual Data Point Visualization

* **Scatterplot**\ : This is equivalent to a scatterplot for the scatterplot with two numeric variables, but data points are aligned in accordance with the given x-axis.

**95% Confidence Interval**\ : This shows 95% confidence interval for each group of a category variable on x-axis. A box placed upon the data points of a group represents the confidence interval of a sample mean for the group. When you click the box, you will be shown another box that is long and horizontal that you can use to compare with other groups' confidence interval boxes. You can imagine conducting an eye-level two-sample t-test.

**Distribution**\ : This is one of the two ways to visualize how data points are distributed across y-axis by each group of a categorical variable on x-axis. Length of a line represents the number of data points lying upon the line.

**Heatmap**\ : Heatmap, just as

Distribution does, represents the distribution of data points but with gradient colors. Note that you can change the color scale with the color scale icon

on top right of HEARTCOUNT.

Grouped Data Point Visualization

* You may create subgroups based on the categorical variable on the x-axis by using subgroup tab. After so, Smart Plot will provide a set of available options for grouped data point visualization.

* There are five metrics you may choose to put on a y-axis. * Average of a numeric variable * Sum of a numeric variable * Record count: how many records are in each group * Global ratio: how much proportion each group takes within the given dataset * Local ratio: how much proportion each subgroup takes within the each group. This is useful when you divide each group into another subgroup. * Here are the list of available visualization options.
*

**Bar chart**\ : This is a simple bar chart where heights are proportional to the given metric on a Y-axis.

* You may use other variable to create subgroups within each group instead of grouping the entire group. Each bar on the x-axis will be divided into multiple subgroup bars.

**Stacked Bar Chart**\ : Another way to use the subgroup feature is with a stacked bar chart. Each bar on the x-axis will be divided into multiple subgroup bars, and the subgroup bars will be stacked so that the original bar can be better understood.

**Stacked Area Chart**\ : A stacked area chart is similar to the stacked bar chart, except that subgroup bars are connected together so that the levels/values of each subgroup vary over the given x-axis variable can be better understood. Below is displayed how easily local ratio of subgroups within each category can be calculated.

**Line Chart**\ **:** This is a traditional line chart where seasonal patterns and trends can be explored if your dataset contains a temporal(time, date, etc) variable.

{% endtab %} {% tab title="Time Series on X & Numeric on Y" %} You also can put a time-series variable on x-axis.

See in Detail

* If your dataset contains a time series variable, you can map it on the x-axis. Smart Plot will then display a line chart.

* Smart Plot will generate a monthly time interval for you. To switch to a different time interval, such as weekly or yearly, use the small clock-like icon

next to the x-axis tab.

* You can use automatically derived time series variables such as days of week, day type(weekdays or weekend), or so to create a plot. This will enhance your understanding of the relationship between the time series variable and others.

* You can check how the variable on the y-axis has evolved over the x-axis by using the trend line feature

* Smart Plot also supports two simple forecasting models

, namely the max entropy model and the least square method.

{% endtab %} {% endtabs %} ### - Additional Features #### **Facetting** Smart Plot offers a facet feature. It allows you to divide a single plot into multiple charts based on a facet variable in order to better understand the relationship between the x and y variables. You can also use every other feature in Smart Plot within each facet plot.

{% hint style="info" %} Categorical variables with fewer than 11 groups can currently be used as a facet variable. {% endhint %} {% hint style="info" %} To get free access to HEARTCOUNT Public, [Click Here](https://www.heartcount.io/login?lang=en) {% endhint %} # Drill-Down ## :question:What It Can Do for You

Drill-Down allows you to easily compute, summarize, and analyze data, allowing you to find similarities, hidden patterns, and overall trends of key metrics in your data. With Drill Down, you can divide it into multiple dimensions within dimensions by simply selecting which variable you want to segment the data (this is usually called a pivot table). Drill Down offers you three different ways to represent the created drilled-down data so that you have much deeper understanding of the dataset. ## :page\_with\_curl:How to Use There are three different types of visualization you may utilize in Drill-Down menu. They all serve one purpose: to help you understand how key metric differs within given segments. In addition, Drill-Down provides SMART DISCOVERY section where an automatically discovered combination of variables that maximize given key performance index. How is the easiest part of using HEARTCOUNT's Drill-down menu, just as the other functions of HEARTCOUNT are. You just have to simply...\ \ \ **1. Select a variable of your interest as KPI**
* **In this step, there are four options you can choose.**\ **1) Count of rows in the dataset:** \ *"Which region has created the most number of transactions?"*\ **2) Ratio of given segment within other segments:**\ *"How much ratio do technology products make up about?"*\ **3) Average of the variable:** \ *"Which category has recorded the highest average profit?"*\ **4) Sum of the variable:** \ *"Which state has been the most profitable?"*\

\ **2. Select variables to segment the data and click Run**

* You can freely choose any of variables in the dataset as segment variables except the one selected as the key metric variable previously. * If you choose numeric variable as a segment variable, HEARTCOUNT will automatically pick up derived bins of the selected variable. You may switch this to derived percentile variable. * It is possible to choose multiple segment variables. In this case, Drill-Down creates segments within segments in the order of the chosen segment variables.

\ **3. Choose a type of visualization** Choose which visualization method you wish to use from the three different types of visualizations, which are described below. ### - Types of Visualizations ####

Drill-Down

\ Drill-Down is a main function of Drill-Down menu. With given key metric variable and segment variable, Drill-Down shows you how average/sum of key metric value varies within each segment with arrows and numbers. Length of an arrow shows a relative size/amount of the metric for each segment. * On top right of the drill-down table is displayed an overall average for average metric, grand total sum for sum metric, overall average ratio for ratio metric, and grand total count for count metric. * By clicking the order icon

upon the top left of the drill-down table, you can decide in which order you wish to arrange the table. *

: sort by value order (ascending) *

: sort by value order (descending) *

: sort by alphabetic order (ascending) *

: sort by alphabetic order (descending) * Upon the top right of the drill-down table is where you can set the record count threshold. \

* If a segment consists of less number of records, it will be excluded in the table. ### ####

Tree Map Tree Map provides a different visual approach to understand how the metric varies within segments. Tree Map is like opening boxes within boxes and size of each box represents the value of the box itself. The first segment variable you choose will be the outermost box(root of the tree, first layer), and the second segment variable will consist of the second layer, which is why this type of visualization is called Tree Map.

To construct a Tree Map visualization, you must first select the key metric variable and segment variables in the order in which you want to segment your data, just as you would do in the Drill-Down visualization tab. The first layer of the tree map will appear after you click the Run button (outermost box, root layer of the tree). If one of the first layer segments piques your interest, simply click the box to examine how the metrics for the second layer segments inside the segment of interest fluctuate.

# Small Multiples ## :question:What It Can Do for You Small Multiples gives a bird's-eye view of a single numeric variable over all other variables in the dataset. Small Multiples enables you to instantly understand the overall relationship between the selected variable and the other variables by showing you the relevant charts. ## :page\_with\_curl:How to Use To use Small Multiples, all you have to do is to just select one of the numeric variables. Then, Small Multiples will generate multiple charts of two types that reveals you the relationship between the selected variable and the others. One is a scatterplot for a numeric variable, the other is a bar chart for a categorical variable. For instance, if you want to understand the relationship between a key metric variable, such as profit, and all other variables, simply pick profit as a y variable in Small Multiples.

If you simply want to look at the link between profit and all other factors in Texas, you can use the Record Filter feature just as below.

All simplified scatterplots that show the relationship between profit and numeric variables are located on the top,
whereas bar charts for categorical variables are located on the bottom.

From there, perhaps the relationship between profit and product category in Texas catches your attention. Then, simply move your cursor over the bar chart titled as "Category" and click the Smart Plot symbol that appears in the upper right corner of the chart.

# Smart Ranking Comparison between groups done easily ## :question:What It Can Do for You Our Smart Ranking feature enables you to quickly generate multiple bar charts based on the entire set of numerical values(sales, weights, costs, etc) in a given dataset over a categorical variable(groups, ids, types, product categories, etc) you choose. This gives you a chance to take a broad look of the dataset and makes it simple to understand which group generally has greater numeric values. ## :page\_with\_curl:How to Use By simply selecting single categorical variable(e.g. product category, region, ...), HEARTCOUNT displays multiple bar charts of all the numeric variables(profit, quantity, ...) by each value of the selected categorical variable. Let's see how it works in practice from Superstore dataset. * Say you want to how each numeric value varies on average by each product category. The only action you have to take is simply choose Category as a ranking variable. * Then, HEARTCOUNT will produce multiple bar charts with bars ordered in decreasing order on each chart. With the created bar charts, you will have a overview of how each numeric variable differs by each category.
* Furnitures, for example, which are shown by blue bar charts, have averaged medium sales but the lowest profit, which is exceptional. However, if you look at the discount bar chart, you can see that on average, furniture was sold to clients at a larger discount rate, which may have influenced average furniture profit.

### * You may highlight one of the bars (categories) and make it more apparent and distinct by either using the search tool or by clicking on one of the bars. This will lead to a better comprehension of the selected category.

* If you want to examine if this relationship between category and each metric changes inside a certain state, like California, you may use the HEARTCOUNT record filter to exclude all other states except California. * Even with the largest discount rate, it appears that technology-related items earned the best profit and sales in California. * Meanwhile, the quantity of tech items was the biggest, but the record count (number of transactions) was the lowest. This might imply that a small number of clients in California acquired a large number of tech goods in a few transactions.

* If you want to learn more about the link between product category and profit, move your cursor over the profit bar chart and click the Smart Plot symbol that appears in the upper right corner of the chart. HEARTCOUNT will then switch to Smart Plot, with category on the x-axis and profit on the y-axis.

# Smart Facets ## :question:What It Can Do for You Smart Facets is a very useful feature of HERATCOUNT when you want to better understand the relationship between two variables by creating multiple facets with the selected facet variable. For example, if you want to create a bar chart about sales with respect to ship mode from the Superstore dataset and how it varies over the regions, all you have to do is select 'region' as facet variable in Smart Facets.

So, what's the difference between Smart Plot's facet function and Smart Facets? In Smart Plot, categorical variables with less than 11 levels can be used as a facet variable for +++. On contrary, Smart Facets allows you to utilize any categorical variable with more than 11 levels. ## :page\_with\_curl: How to Use As straightforwardly as it could be, to use Smart Facet, you just decide which variable in your dataset you would like to use as x,y variable, and then choose another variable to use as a facet variable, which follows typical visualization grammar just as the other menus of HEARTCOUNT do. Let's take a look at how it works in the Superstore dataset. * Say you want to understand how the quantity of sales varies over product categories and how it changes over different regions. All you have to do is to just simply choose Category as a X-axis variable, Quantity as a Y-axis variable, and Region as Facet variable. Then, Smart Facets automatically create multiple bar charts, showing the relationship between Product Category and Quantity over different Region.

{% hint style="info" %} Note that both categorical and numeric variables are allowed, whereas only numeric variables are allowed for y-axis variable. {% endhint %} {% hint style="info" %} Notice a small icon on the top-right for ordering the facet. Currently, facets are ordered in a way that reflects statistic order, which is +++ {% endhint %} * If you mouse over one of the bars, which is one of the product categories, it will show you how much is the average quantity(=height of the bar) for the category. Let's say how many furnitures were sold to each customer on the west on average. If you mouse over a blue bar on the West facet, it will show the mean quantity of furniture sold on the west.

* You may want to further investigate on the relationship between the quantity and the category over different states in the West. To do so, you have to use the record filter function first, which is on the of the page, where you can filter out all the other regions(Central, East, South). Then, choose State as a facet variable.

Montana sold more than the other states on average,
and the quantity of furniture sold was the most on average among the product categories in the West.

# Smart Search Smart Search empowers you with a visual exploration tool, letting you sift through similarities based on your search criteria. ## **1. What is Smart Search?** Smart Search allows you to compare similarities between specific records. Not only can you visualize the distribution of similarities based on your search conditions, but you can also quantify them with a numeric "similarity" score. The closer a record's similarity score aligns with your criteria, the nearer it will appear to the center in a concentric circle representation. ## **2. How to Use Smart Search** The image below showcases the Smart Search interface, displaying an entire staff clustered by their \[Business Location]. The search criteria are set to \[Performance Score: 77.1\~94.1] and \[Unused Vacation: 12.3\~31]. By clicking on the individual points, you can access detailed data.

* **Search Criteria**: Choose the variables and their values to compute similarities. Click on the dropdown menu to select a variable, then either input its value directly or use the mouse to select a range. * **Cluster**: Select the variable to cluster individual items based on their proximity. In the given example, staff members are clustered by \[Business Location], showcasing groups of employees working at the same location. * **Color**: Differentiate items on the grid using a selected color variable. While the default color is set based on the clustering variable, as in the example, you have the freedom to choose another variable if desired. * **Size**: This determines the relative size of individual items (points). * **Filtering**: Use this to filter items on display by any variable. Click on the dropdown, choose your desired variable, and then define its range. If you click on the `table view` icon located at the top right of the image below, you'll switch to a table format, revealing detailed information about each record.

In the provided example, cells corresponding to \[Performance Score] and \[Unused Vacation] are highlighted based on the search criteria.

When you set multiple search criteria, records are primarily sorted based on their similarity score (0\~100%). The color intensity varies with the score - the closer a record aligns with the search criteria, the deeper its color. Conversely, records with low similarity scores are represented in lighter shades.

In essence, Smart Search offers a visually intuitive way to explore data distributions, colors, sizes, and more, ensuring you find the most relevant data points based on your criteria. # Dashboard Freely set key metrics and automatically find patterns in your data that aren't immediately obvious. ## 1. **What is Dashboard?** The dashboard allows you to view basic summary information about a campaign, such as the number of records (cells in Excel) and the average of KPIs set by the user.

> 💡 The 'K' stands for 1,000. ## **2. KPI Filtering** Clicking on the ▼ button next to the KPI on the right side of the dashboard will display the KPI items you've set. By selecting individual items or bulk selecting and then clicking '`Apply`', the chosen variables will be reflected on the dashboard. You can also uncheck the checkbox to remove that KPI from the dashboard. Note that it's only removed from the dashboard, not from the KPIs. :) If you want to delete or modify a KPI, please use the KPIs Management menu.

## **3. KPI Distribution by Specific Variable** Select a specific variable and click on '`Analyze`' to view the distribution of the KPI for that variable in a [drill-down](https://public.heartcount.io/visual-discovery/drill-down) format.

## **4. Detailed Analysis of KPI** Hovering over a KPI at the top of the dashboard will display a menu window. For numerical variables, this menu allows you to navigate to the detailed analysis page for that KPI ([Drill-Down](https://public.heartcount.io/visual-discovery/drill-down), [Driver Analysis](https://public.heartcount.io/advanced-analysis/driver-analysis), [Micro-Segmentation](https://public.heartcount.io/advanced-analysis/micro-segmentation), etc.). For categorical variables, the menu provides options to navigate to the Micro-Segmentation or Drill-Down. Click on the desired analysis icon to go to that analysis page.

## **5. SMART DISCOVERY** SMART DISCOVERY is a feature that automatically detects patterns for the set KPIs. The Smart Discovery function is available in both the dashboard and [drill-down](https://public.heartcount.io/visual-discovery/drill-down).

It provides an overview of patterns and significant results applied to KPIs across all HEARTCOUNT features, such as Drill-Down, Driver Analysis, and Micro-Segmentation. * **Drill-Down I** informs you of the **variable** that most significantly impacts a specific KPI. By clicking on a KPI from the list at the bottom, you can quickly check the variable and value that indicates the highest value for each KPI. * **Drill-Down II** tells you about the **combination of variables** that most significantly impact a specific KPI. By clicking on a KPI from the list at the bottom, you can quickly identify the combination of two variables that contribute the most to each KPI.

Clicking on the `graph` icon to the right of the drill-down will display the distribution of that variable's KPI at the bottom of the dashboard in a drill-down format. # Driver Analysis Understand the factors that influence KPIs. Through the Driver Analysis feature, you can easily identify the crucial factors influencing the KPIs you want to explore. ## 1. **What is Driver Analysis?** {% hint style="info" %} Driver Analysis is possible only for numerical KPIs. If you haven't registered a numerical KPI yet, go to KPIs Management and register one! {% endhint %} There are times when we want to know which factors influence sales the most. By using the **Driver Analysis** feature, you can determine which factors have the most influence on user-defined KPIs, such as sales. Driver Analysis is a regression analysis function that finds the most important factors (variables) for the chosen KPI.

Learn more about HEARTCOUNT's regression analysis

Regression analysis is a data analysis method used in statistics to estimate and explain relationships between variables. By setting an influencing variable and an influenced variable (KPI), you can understand the relationship between the two. Analyzing the relationship between one variable and a KPI is called simple regression analysis, and when seeking to clarify the relationship between multiple variables and a KPI, it's called multiple regression analysis. In HEARTCOUNT, we use both simple and multiple regression analysis methods to present significant results in order of importance. 👉 Click here to learn more about regression analysis

## **2. How to Use Driver Analysis?** Here's a basic description of the Driver Analysis interface. Each section of the interface has the following functionalities:

1. **Filter** This has the same function as the dashboard filter. You can select or exclude desired variables to view Driver Analysis results. By clicking on the `Variables` tab to the right of the filter, you can include or exclude derived (automatically generated) variables for the analysis. Derived (automatically generated) variables are HEARTCOUNT's unique feature that auto-processes numerical variables into categories to discover non-linear patterns. Click here for a detailed explanation of this feature. 2. **Variable Selection & Auto-play** * **\[All]** allows you to see the results integrated from simple and multiple regression analysis. Simple regression can be viewed under **\[1 Variable]**, and the results of multiple regression analysis can be detailed under **\[2 Variables]**. * The factor analysis results explain the difference of the KPI in order of Adjusted $$R^2$$, showing which variable can explain how much (%) of the KPI's variance and the corresponding visualization. When you click the play ▶︎ button, a video like the one below will appear.

3. $$R^2$$ $$R^2$$ represents the coefficient of determination, a statistic that reflects the **proportion of the variance in the dependent variable (KPI) that can be explained by the independent variable**. For example, if the selected KPI, \[Sales], is best explained by $$R^2$$ being 0.56 with the highest values from the variables \[Profit\_bin] and \[Sub-category], it means that 56% of the variance in \[Sales] can be explained by \[Profit\_bin] and \[Sub-category].

4. **Adjusted** $$R^2$$ Adjusted $$R^2$$ denotes the adjusted coefficient of determination, which modifies the $$R^2$$ value considering the number of records and variables used in the analysis. This metric **compensates for the inherent issue of the coefficient of determination increasing indiscriminately** with an increase in independent variables. {% hint style="info" %} When using Driver Analysis with two variables, it's essential to interpret both the coefficient of determination and the adjusted coefficient of determination. {% endhint %} 5. **P-Value** The P-Value indicates the probability that the analysis results occurred by chance. Typically, if the P-Value is **less than 0.05**, it's considered **"statistically significant."** Conversely, a P-Value exceeding 0.05 suggests a higher likelihood of randomness, and the relationship is deemed **"not statistically significant."** 6. **Records** The number of records used in the analysis (excluding records with irrelevant values). 7. **Smart Link** Use the Smart Link to move to the [Drill-Down](https://public.heartcount.io/visual-discovery/drill-down) and [Explainer](https://public.heartcount.io/advanced-analysis/explainer) page for the specific variable. 8. **Visualization Screen & Control Window** The topmost result in Driver Analysis is visually displayed on the visualization screen. You can visually verify the results through scatter plots, bar graphs, etc. For a detailed description of the visualization screen, click [here](https://public.heartcount.io/visual-discovery/smart-plot), and for the control window, click here. # Explainer Understand the drivers influencing the KPI. With the Explainer feature, you can easily understand the Driver Analysis results for the chosen KPI in everyday language. ## 1. **What is Explainer?** The Explainer is a feature that describes the Driver Analysis results for a specific KPI in natural language, making it easier for users to understand. It removes complex statistical concepts, allowing users to quickly and easily understand the factors influencing the KPI through visualization. {% hint style="info" %} **Natural language** refers to the language humans use daily, contrasting with constructed languages created for specific intentions and purposes in fields like computing and mathematics. {% endhint %} ## **2. Screen Composition & How to Use** Here's a description of the basic screen composition of the Explainer. Each section of the screen below has the following functionalities:

#### (1) **Basic Usage** It lists the factors that have the most significant impact on the KPI. By clicking on the 🔁 icon, you can change the order of the variables. By pressing the right arrow (>), you can see the explanation for the next ranked factor. It describes in plain language the combination of conditions that are higher or lower than the average of the selected KPI.

#### (2) **Generate New Narrative** It provides an explanation from a **different perspective** regarding the given factors. For instance, even if the previously set KPI, \[Sales], indicated conditions that were above or below average, clicking the "`Generate New Narrative`" button will show the combination of conditions that contributed the most or least to the total \[Sales]. ![](https://1130022100-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FHmGL0fwQrDEm9mSq6FQN%2Fuploads%2FxSOHfL59meNuk9aL7izj%2FScreenshot%202023-09-04%20at%202.12.28%20PM.png?alt=media\&token=324a030a-eef3-4d2c-ad08-7a1dc7ac8a7e)

#### (3) **Change Visualization** Clicking the "`Change Visualization`" button will display a different type of graph in the bottom right visualization window.

#### (4) **Visualization Window** This feature visualizes the driver analysis results on the left in the form of graphs. The darker the color (blue), the higher the explanatory power for the KPI, and the lighter the color (yellow), the lower the explanatory power. By clicking on the graph, you can check the analysis results for the corresponding factor through the description on the left.

# Difference Analysis Understand the differences in characteristics between two groups. The Difference Analysis feature allows you to discern the statistical differences between any two groups. ## 1. **What is Difference Analysis?** Every company experiences a scenario where one group performs exceptionally while another doesn't. **😁 Wouldn't it be great if you could automatically identify the factors causing these differences?** HEARTCOUNT offers a Difference Analysis feature that can be utilized when you're curious about the statistical differences between any two groups. Through this process of understanding the differences between the two groups, you can pinpoint which variables require attention and which need improvement. ## **2. Setting up Difference Analysis** 1. Click on the `+` icons on both the left and right to set up two arbitrary groups (Group A and Group B) whose statistical characteristics you wish to compare. (The example at the top compares conditions between 'Group A\_with lower profits' and 'Group B\_with higher profits'). 2. Click the '`compare`' button on the top right. 3. Variables showing the **most significant statistical differences between the two groups** set in the analysis conditions are displayed in descending order. 4. If you want to specify the groups in more detail, click on the `+` icon to add specific variables using the AND condition for comparison.

> *\*Refer to the screen below:* **Group A\_**with lower profits and located in Arizona, California, and Georgia vs. **Group B**\_with higher profits and located in New York, North Carolina, Ohio, and Oregon.

## **3. Overall Explanation of Difference Analysis**

1. **Variable Name**: The name of the variable explaining the differences between the two groups and the statistical difference are shown in percentages. In the illustration above, the second variable, 'Profit', with a value of 72.41%, means that when comparing the 'Profit' distribution between the two groups, if the entire distribution area is considered 100, then the **non-overlapping distribution area** is 72.41%. Put differently, *the larger this number, the better it represents the statistical difference between the two groups*. 2. **Chart**: This visually represents the statistical differences between the two groups. Group B (Profit=0\~2.2K) indicates a higher profit group than Group A (Profit=2.3K\~22.6K). 3. **Statistical Characteristics of Group A/B**: This displays the variable value (range) where a specific group has a relatively higher ratio than the other group. \ *\*In the example screen above, the 'Profit' in the 484.9\~3919.99 range shows the 'higher profit group' exists at a higher ratio than the 'lower profit group'.* # Micro-Segmentation Discover patterns to optimize KPI With the Micro-Segmentation feature, you can discover actionable rules to optimize any KPI. The KPI mentioned here refers to the target variable of the analysis. Thus ## 1. **What is** Micro-Segmentation**?** Professionals always ponder about optimizing their tasks. Imagine a marketing executive in a company who, during a campaign, identifies rules to segregate groups that are likely to respond enthusiastically and those that won't. Such insights would allow the company to efficiently target the potential group. Wouldn't it be wonderful if one could **automatically identify ways to optimize desired goals?** HEARTCOUNT provides Micro-Segmentation feature, which discovers rules to optimize KPIs automatically, recommending conditions for optimal execution. This feature leverages the *decision tree algorithm*, a prominent machine learning technique used to classify two different groups.

What is a Decision Tree?

A decision tree algorithm designs models to classify distinct groups. Simply put, it delineates spaces to **group similar entities together.** The significance of this algorithm is that it formulates **optimal classification rules** using as few variables as possible, considering real-world datasets often comprise numerous variables. * *Want to know more about Decision Trees? (Click)*

## **2. Utilizing Micro-Segmentation** The example below illustrates the process of automatically discovering rules to optimize profits in a discount mart. Let's explore how to find optimization rules for a particular KPI using Micro-Segmentation. ### (1) **Finding Optimal Rules** Begin by selecting the variable you wish to optimize (e.g., \[Profit]). Clicking the '`analyze`' button on the top right will automatically suggest rules to classify the **top 20%** and **bottom 20%** in the table.

💡 Micro-segmentation provides 'optimization rules' for **numeric target variables** by automatically segregating them into top/bottom groups. For **categorical variables**, classification rules are offered, which can be set as follows.

#### **(Optional)** At its core, micro-segmentation aims to differentiate between two distinct target groups and establish a set of 'rules' for their classification. While we provide a default split of the top 20% and bottom 20% groups, users have the flexibility to define their own target groups. Simply use the `cogwheel ⚙️` icon to **adjust the range for the groups you're interested in.** You can easily slide the range with your mouse or input specific values.

#### **Dive Deep into the Table Details** Each entry in our table comes with its unique information. As a quick guide, entries are ranked in descending order based on their target ratio.

Let's break them down: 1. **Purity**: Represents the proportion of a specific target group within a larger segment, giving insights into the accuracy of a classification rule. For instance, consider how the subgroup with 'less than 12% discount rate and in the top 20% profit' fits into the broader group having 'a discount rate less than 12%.' 2. **Target Ratio**: It's the proportion of segments that adhere to a particular classification rule within the main target group. For example, think of how the segment with 'a discount rate less than 12%' compares to the larger segment 'in the top 20% of profits.' 3. **Average KPI Difference**: This metric shows the gap between the average value of a specific segment and the overall target variable. 4. **Smart Search (▽)**: Want more detailed insights? Click on the ▽ next to each row. This action will lead you to our 'Smart Search' feature, allowing you to dive deeper into individual data records for each rule. \ 👉 Intrigued by Smart Search? Give it a click! #### **A Glimpse into the Predictive Model's Summary** For a concise overview, click on the downward arrow (∨) highlighted in a red box in the provided visual.

This summary gives you insights into either the default target groups or any custom ones you've crafted.

> 🚨 It's crucial to note that the summary's target ratio varies from the individual segments mentioned previously. \ > \ > Although the overall target remains constant, individual segment targets might differ. In this summary, we amalgamate all the derived rule segments and then divide by the total target. To illustrate, if the top 20% profit group stands at a size of 1.64k, and the total derived segments sum up to 1.53k, it signifies that an impressive 93.6% of the top 20% profit group aligns with significant classification rules. ### (2) **Visualization Graphs** Users can understand the provided rules more intuitively through visualized graphs. The image below showcases the visualization of the rule at the top of the table, which is based on the **'Profit Optimization'** criterion. It specifically highlights the rule where Sales is more or equal to 86.19, the discount rate is less than 0.12, State is Alabama, Arkansas, and so on.

1. **Class**: Records classified under the condition of `State = [Alabama, Arkansas, ..., Wyoming]`, `86.19 <= Sales`, `Discount Rate < 0.12` predominantly belong to the Top 20% Profit group. 2. **Size**: The number of records satisfying the condition (1,449 records). 3. **Purity**: Among the segment size of 1,449, 1,427 (98.48%) belong to the Top 20% Profit group. A higher purity signifies greater reliability in the analysis outcome. In essence, it's a metric indicating the precision of the classification rule. 4. **Target Ratio**: Out of the total records in the Top 20% Profit group (2,000), 1,427 (71.35%) fall under this segment. 5. **Average KPI Difference**: The difference between the average KPI of the records within this segment and the overall average KPI. 6. **Segment Classification Rule**: Detailed representation of the classification rule presented in the first table in the image above. 7. **Visualization Graph**: A graphical representation of the derived profit optimization condition, with the blue region highlighting the classification rule's domain. 8. **Decision Tree**: Clicking the icon displays the classification rules in a tree structure.

### (3) **Exporting as a File** * Clicking on `SQL` will allow users to view the prediction model as a query.

* Clicking the table icon will save the prediction model as a CSV or XLSX file.