> For the complete documentation index, see [llms.txt](https://flapjax.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://flapjax.gitbook.io/docs/reporting/data-sources-and-metrics.md). # Data Sources & Metrics ### Data Sources Every report starts with a data source, the pool of records it analyses. * **People**, your contact database. It holds built-in fields such as External ID, First Name, Last Name, Email and Mobile Number, plus every custom attribute you added. * **Stacks**, any of your custom data tables. Each stack carries its own attributes. * **Lists**, the items of one [list](/docs/your-data/lists.md). The fields available are the list's own attributes plus the parent entity's fields. That lets you report on the working state you keep in a list, such as stage, owner or priority, rather than on the underlying person or record alone. *** ### Filtering Add filter rules after choosing a data source to narrow which records count. The filter uses the same rule builder as segments: choose a field, an operator and a value. Add as many rules as you need and combine them with AND or OR. Click **Add filter** beneath the data source, build your conditions, then click **Apply**. The report then counts matching records alone. *** ### Target Entities Insight reports run cross-entity analysis through the **Target** setting. You start with one data source and measure a related entity. Take a Stack of orders as your data source. Target the linked People and you count how many unique customers placed orders. Start from People instead and you can target a related Stack to aggregate values from linked records. **Available targets depend on your data source:** * From **People**, target any Stack carrying a person attribute that links back to People. * From a **Stack**, target People, where the stack has a person attribute, or another Stack with a relationship attribute linking to the source stack. * From a **List**, target nothing. A list report aggregates its own items, so the Target setting stays out of the way. A selected target switches the metric options to the target's fields rather than the source's. You can add separate filter rules to the target entity too. *** ### Aggregation Metrics The metric sets what calculation runs on the data. The options follow the field type and the report type. #### Numeric Fields For integer and decimal fields: | Metric | Description | | -------------- | ------------------------------------------------------------- | | **Count** | Number of records | | **Sum** | Total of all values | | **Average** | Mean of all values | | **Min** | Smallest value | | **Max** | Largest value | | **P10 to P99** | Percentile values: 10th, 25th, Median, 75th, 90th, 95th, 99th | {% hint style="info" %} **Note:** percentiles cover Insight reports alone. Historical reports support Count, Sum, Average, Min and Max. {% endhint %} #### Boolean Fields For true/false fields, in Insight reports alone: | Metric | Description | | ----------------- | ------------------------------------------ | | **True Count** | Number of records where the value is true | | **False Count** | Number of records where the value is false | | **Empty Count** | Number of records with no value | | **Percent True** | Percentage of records that are true | | **Percent False** | Percentage of records that are false | #### Date Fields For date and timestamp fields, in Insight reports alone: | Metric | Description | | ------------------- | ------------------------------------------------ | | **Earliest** | The oldest date in the dataset | | **Latest** | The most recent date | | **Non-empty Count** | Number of records that have a date value | | **Range (seconds)** | The span between earliest and latest, in seconds | | **Range (days)** | The span between earliest and latest, in days | *** ### Field-Type Metric Catalogue (Aggregate Records) The metrics above are what the **report builder** offers for aggregating a numeric field or counting records. The **Aggregate Records** automation action goes further. It, and the record-metrics engine behind it, can summarise a field of *any* data type and write the result back into your automation. The metric set available follows the field's data type. Use this catalogue as you build aggregation actions, or as a reference for the summaries the platform can compute per type. #### Text, Select, Multi-select, Labels, and other text-like fields This covers text, select, multi-select, labels, country, currency, source type, label category, mobile number and mobile prefix fields. Multi-select and label values get flattened, so each individual value counts. | Metric | Description | | ------------------ | ---------------------------------- | | **Most Frequent** | The value that appears most often | | **Least Frequent** | The value that appears least often | | **Distinct Count** | Number of unique non-empty values | | **Empty Count** | Number of records with no value | #### Person fields For a person-reference field. | Metric | Description | | ------------------ | ------------------------------------ | | **Distinct Count** | Number of unique people referenced | | **Empty Count** | Number of records with no person set | #### Stack (single-relation) fields For a field that links to a record in another stack. | Metric | Description | | ------------------ | ---------------------------------------- | | **Distinct Count** | Number of unique linked records | | **Empty Count** | Number of records with no link set | | **Most Frequent** | The linked record referenced most often | | **Least Frequent** | The linked record referenced least often | #### Numeric fields (extended) Integer, decimal and time-of-day fields support the full statistical set, well beyond Sum, Average, Min and Max. | Metric | Description | | ---------------------------------- | -------------------------------------------------------- | | **Count (Non-Empty)** | Number of records that have a numeric value | | **Sum** | Total of all values | | **Average / Mean** | Mean of all values | | **Minimum / Maximum** | Smallest and largest value | | **Percentiles** | 10th, 25th, 50th, 75th and 99th percentile | | **Median** | The middle value, the 50th percentile | | **Interquartile Range (IQR)** | The spread between the 25th and 75th percentiles | | **Std Dev (Population / Sample)** | Standard deviation | | **Variance (Population / Sample)** | Variance | | **Median Absolute Deviation** | A spread measure around the median that resists outliers | | **Skewness** | How asymmetric the distribution is | | **Kurtosis** | How heavy-tailed the distribution is | #### Boolean fields (extended) | Metric | Description | | ---------------- | ---------------------------------------------- | | **True Count** | Number of records where the value is true | | **False Count** | Number of records where the value is false | | **Empty Count** | Number of records with no value | | **Percent True** | Percentage of non-empty records that are true | | **Most Common** | Whether true or false is the more common value | #### Date and Date-time fields (extended) Date and date-time fields support a rich set of temporal summaries. Date-time fields add **Most Common Hour** and **Nighttime Percentage** on top. | Metric | Description | | ------------------------------------------- | --------------------------------------------------------------------------------------------- | | **Earliest / Latest Date** | Oldest and most recent value | | **Date Range (Days)** | Time span between earliest and latest, in days | | **Midpoint Date** | The date halfway between earliest and latest | | **Time Since Earliest** | How long ago the earliest value was, in seconds, minutes, hours, days, weeks, months or years | | **Time Since Latest** | How long ago the latest value was, in the same units | | **Most Common Day of Week** | The weekday values fall on most often | | **Most Common Month** | The month values fall in most often | | **Weekday / Weekend Percentage** | Share of values falling on weekdays vs weekends | | **Regularity Index** | How evenly spaced the values are, where higher means more regular intervals | | **Interval: Avg / Median / Min / Max** | Gap between consecutive values, in seconds, minutes, hours, days or weeks | | **Most Common Hour** *(date-time only)* | The hour of day values occur most often | | **Nighttime Percentage** *(date-time only)* | Share of values occurring overnight, midnight to 5am | {% hint style="info" %} The field's data type decides the exact metrics you see as you build the action. Multi-relation and label fields cannot serve as the grouping dimension in reports, and their values still count as text-like metrics here. {% endhint %} *** ### Grouping A **Grouped by** attribute breaks your data into categories. Group people by country, for instance, and you get a bar per country. Most attribute types work for grouping. Multi-relation and label fields do not. With no grouping applied, an Insight report shows a single metric. Add grouping and the result becomes a chart. **Include empty values** decides whether records with no value for the grouping attribute form their own group. *** ### Segmentation A **Segment by** attribute cross-tabulates your data. Segmentation adds a second dimension, splitting each group by the segment attribute's values. In Insight reports, segmentation forces a Bar chart, with each segment as its own coloured bar, grouped or stacked. In Historical reports, it produces one line per segment value on the time-series chart. **Include empty values** works on the segment attribute independently of the grouping attribute. *** ### Time Range Historical reports need a date range. The series buckets by each entity's **creation date**, in UTC, so the report shows how much data was *created* in each interval. Pick a range from the preset bar: | Preset | Meaning | | ----------------------------------------- | ---------------------------------------------------------------------------------------------- | | **7D** | The last week | | **1M** | The last month | | **3M** | The last quarter | | **6M** | The last six months | | **12M** | The last twelve months | | **Week / Month / Quarter / Year to date** | From the start of the current period until now. *Month to date* on July 28 covers July 1 to 28 | | **Custom** | Any from-and-to range you choose. The picker takes a date **and** a time on both ends | **Rolling and to-date differ.** *1M* is a rolling window of roughly the last 30 days, ending today. *Month to date* resets at each period start. Pick to-date for tracking progress against the current week, month, quarter or year. *** ### Time Granularity Granularity divides the range into **Hourly**, **Daily**, **Weekly** or **Monthly** intervals, one data point each. Flapjax offers the granularities that fit your selected range. Change the range to one that cannot support your current granularity and it auto-corrects to the closest sensible option. | Range | Granularities offered | Default | | -------------------- | ------------------------------------------------------------------------------------------------------------------------------------------ | ------- | | 7D / Week to date | Daily | Daily | | 1M | Daily, Weekly | Weekly | | Month to date | Daily, Weekly | Daily | | 3M / Quarter to date | Weekly, Monthly | Weekly | | 6M | Weekly, Monthly | Weekly | | 12M | Monthly | Monthly | | Year to date | Weekly, Monthly | Monthly | | Custom | Follows the span. Hourly up to 2 days, Daily from 1 day to roughly 3 months, Weekly from 1 week to a year, Monthly from roughly 1 month up | Daily | One rule of thumb sits behind those limits. A chart stays readable up to roughly **60 data points**. That is why Hourly over a month, at 720 points, and Daily over a year are both off the table. *** ### What differs on list reports * A list source asks you which list, and the report then covers that list's current members. * No **Target** step exists. List reports aggregate the list's own items. * **Contributing data** works. Click through an aggregated value and the underlying list items open. * **CSV export of contributing data is not available** for list sources. The interface hides the export action, and the API rejects the request.