Best Approaches to Analyzing and Displaying Customer Reviews with Star Ratings in Mixpanel

·Jun 01, 2026 01:01 PM·

(edited)

I'm tracking customer review with Star rating and free text input What is the best approach to analyse and display the results in Mixpanel please provide examples

11 comments

· Sorted by Oldest

kapa.ai

APP

(edited)·

Here's a recommended approach based on Mixpanel best practices:

Rather than creating separate events for each rating value, track one event with rich properties. [single event power] Example event:

Property	Type	Example Value
event name	—	Review Submitted
star_rating	Numeric	4
review_text	String	"Great experience!"
product_id	String	"SKU-123"
review_category	String	"Delivery"
platform	String	"iOS"

⚠️ Important: Keep star_rating as a Numeric type so you can calculate averages, medians, and percentiles in Mixpanel. [property data types]

⚠️ For review_text, be mindful that strings are truncated at 255 characters, and free-text fields may contain PII. Consider whether you need to send this to Mixpanel at all — tagging/categorising the text before sending (e.g. review_category) is often more analytically useful. [import events]

Use the Insights report to visualise your review data: [Insights overview]

Behavior: Review Submitted
Measurement: Aggregate → Average of star_rating
Visualization: Line chart → shows rating trend over time

Behavior: Review Submitted
Measurement: Total Events
Breakdown: star_rating
Visualization: Bar chart → shows how many 1★, 2★, 3★, 4★, 5★ reviews were submitted

Behavior: Review Submitted
Measurement: Average of star_rating
Breakdown: product_id or review_category
Visualization: Bar chart → compare ratings across products/categories

[Insights quick start]

You can group ratings into sentiment buckets directly in the UI without changing your data: [discover insights]

Negative: 1–2 ★
Neutral: 3 ★
Positive: 4–5 ★

This uses Mixpanel's custom buckets feature on the star_rating numeric property.

Filter to only show low-rated reviews (star_rating <= 2) and break down by platform or product_id to find problem areas.
Break down by user properties (e.g. subscription plan, country) to see if certain user segments leave lower ratings. [Insights quick start]

Goal	Report	Key Setting
Rating trend over time	Insights (Line)	Avg of star_rating
Rating distribution	Insights (Bar)	Breakdown by star_rating
Ratings by product	Insights (Bar)	Breakdown by product_id
Sentiment grouping	Insights	Custom buckets on star_rating
Low-rating deep dive	Insights	Filter star_rating ≤ 2

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai what is the ebst practice to display long text results in this kontext?

kapa.ai

APP

(edited)·

The knowledge sources do not contain enough information to describe a dedicated "long text display" feature in Mixpanel for free-text review content. However, here is what the knowledge sources do explicitly recommend:

String properties are truncated at 255 characters. Any review text longer than that will be cut off. [property data types]
Free-text fields may contain unintentional PII (names, emails, etc.) — Mixpanel's privacy guidance explicitly warns: "Don't send unreviewed free-text." [data privacy guide]

Rather than sending raw free-text to Mixpanel, the knowledge sources recommend transforming long strings into structured, analysable properties: [import events]

"Parse the JSON and flatten it into properties... this is much more useful in analysis, as you can filter or breakdown by any key."

Applied to review text, this means categorising or tagging the text before sending it to Mixpanel:

Instead of...	Send this...
review_text: "Delivery was slow and packaging was damaged"	review_theme: ["Delivery", "Packaging"]
review_text: "The app crashes on checkout"	review_theme: "Bug", review_sentiment: "Negative"

This way you can filter, break down, and aggregate review themes meaningfully in Mixpanel reports.

Mixpanel is not designed as a free-text storage or display tool. The best practice is to process your review text externally (e.g. via sentiment analysis or tagging), and send only the structured outputs (categories, sentiment scores, themes) as properties alongside your star_rating. This gives you far richer analytical capability within Mixpanel's reports.

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai how do I bets set up a report that shows the development of the average rating over time

kapa.ai

APP

(edited)·

Follow these steps in the Insights report: [Insights quick start]

1. Select your Event

Click "Select Metrics"
Choose Events → select Review Submitted

2. Change the Measurement

Click on the default "Unique Users" measurement
Select Aggregate Property → Average
Choose star_rating as the property

[measurements] 3. Set the Visualization

Choose a Line chart — this is the best chart type for showing a trend over time [Insights charts]

4. Set your Date Range

Choose a meaningful time window (e.g. last 30, 60, or 90 days)
Set the time interval (daily, weekly, or monthly) depending on your review volume

Enhancement	How
Smooth out spikes	Enable Rolling Average in Analysis Settings [analysis settings]
Compare to a previous period	Use Compare to Past (e.g. vs. last month) [time period comparison]
Break down by product/category	Add a Breakdown on product_id or review_category
Filter to a platform	Add a Filter on platform = iOS / Android

If your review volume is low and daily averages are noisy, the Rolling Average setting is especially useful — it smooths the line by averaging across a sliding window (e.g. 7 days), giving a clearer picture of the true trend. [analysis settings]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai how can I prevent, that periods without active feedback are counted against the avearge

kapa.ai

APP

(edited)·

I'm sorry, but the knowledge sources do not explicitly describe a built-in Mixpanel feature that automatically excludes zero-activity periods from an Aggregate Property Average calculation.

The Average measurement in Insights calculates the average of a numeric property across all instances of the event. [aggregate property] This means:

If no Review Submitted events occurred on a given day, that day simply won't have a data point — it is not counted as a zero in the average calculation.
The risk is more about sparse data causing visual gaps or misleading spikes in the line chart, rather than zero-days dragging the average down.

Problem	Suggested Approach
Noisy spikes on low-volume days	Enable Rolling Average to smooth the trend [analysis settings]
Too many gaps in the line	Switch the time interval from Daily → Weekly or Monthly to aggregate more reviews per data point [date range]
Outlier periods skewing results	Use Compare to Past to visually identify anomalous periods [time period comparison]

Switching to a weekly or monthly interval combined with Rolling Average is likely your best option within Mixpanel's current capabilities to produce a stable, meaningful average rating trend that isn't distorted by low-activity periods. For more precise control (e.g. only averaging days with a minimum number of reviews), this logic would need to be handled outside Mixpanel, such as in a data warehouse or BI tool.

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai ratings can only be 1 through 5 - how can i prevent counting of 0 values on days that don't have feedback at all

kapa.ai

APP

(edited)·

The knowledge sources do not explicitly describe a built-in Mixpanel mechanism to exclude zero-value days from an Aggregate Property Average calculation. However, based on what the knowledge sources do confirm, here is the key clarification:

When using Aggregate Property → Average on star_rating:

Mixpanel only calculates the average across actual instances of the event Review Submitted
Days where no reviews were submitted simply produce no data point — Mixpanel does not insert a 0 value into the average calculation [aggregate property]

So a rating of 0 should not appear unless your tracking code is explicitly sending star_rating = 0 as a property value.

This would most likely mean star_rating = 0 is being tracked in your data. In that case, you can exclude it with an inline filter:

1.
Click the … icon next to your Review Submitted event
2.
Add a filter: star_rating is greater than 0 (or more precisely: star_rating is between 1 and 5)

This ensures only valid ratings are included in the average calculation. [filters]

Scenario	Solution
Days with no reviews	No action needed — Mixpanel won't count them as 0
star_rating = 0 being tracked	Add an inline filter: star_rating > 0
Gaps/spikes from low-volume days	Use Rolling Average to smooth the trend [analysis settings]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
(edited)·
kapa.ai how can i change the Type of properties from string to number, that are sent per integration (ie userflow)

kapa.ai

APP

(edited)·

You can handle this directly in Mixpanel using Typecasting, without needing to change anything in your integration (e.g. Userflow). [typecasting]

1.
Open any report (Insights, Funnels, or Retention)
2.
Add your property as a filter or breakdown
3.
Click the overflow menu (⋯) next to the property
4.
Select "Data Type"
5.
Change the type from String → Number

This forces Mixpanel to interpret the string value as a numeric type, enabling calculations like sum, average, median, and percentile. [typecasting]

Point	Detail
Report-level only	Typecasting applies only within that specific report view — it does not change how the data is stored
String must be numeric	The string value must be interpretable as a number (e.g. "4" → 4). If it cannot be parsed, the value becomes undefined [segmentation expressions]
Not persistent	You need to reapply the typecast each time you build a new report

If star_rating is being sent as a string (e.g. "4") by Userflow:

1.
Add star_rating as your breakdown or measurement property
2.
Typecast it to Number
3.
Then select Aggregate → Average to get your average rating trend

If you want the property to always behave as a number without manually typecasting each time, the best options are:

Fix it at the source — configure Userflow to send star_rating as a numeric value rather than a string, if the integration allows it
Use a Custom Property in Mixpanel to permanently define the transformation [custom properties]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

kapa.ai

APP

(edited)·

Here's a recommended approach based on Mixpanel best practices:

Rather than creating separate events for each rating value, track one event with rich properties. [single event power] Example event:

Property	Type	Example Value
event name	—	Review Submitted
star_rating	Numeric	4
review_text	String	"Great experience!"
product_id	String	"SKU-123"
review_category	String	"Delivery"
platform	String	"iOS"

⚠️ Important: Keep star_rating as a Numeric type so you can calculate averages, medians, and percentiles in Mixpanel. [property data types]

⚠️ For review_text, be mindful that strings are truncated at 255 characters, and free-text fields may contain PII. Consider whether you need to send this to Mixpanel at all — tagging/categorising the text before sending (e.g. review_category) is often more analytically useful. [import events]

Use the Insights report to visualise your review data: [Insights overview]

Behavior: Review Submitted
Measurement: Aggregate → Average of star_rating
Visualization: Line chart → shows rating trend over time

Behavior: Review Submitted
Measurement: Total Events
Breakdown: star_rating
Visualization: Bar chart → shows how many 1★, 2★, 3★, 4★, 5★ reviews were submitted

Behavior: Review Submitted
Measurement: Average of star_rating
Breakdown: product_id or review_category
Visualization: Bar chart → compare ratings across products/categories

[Insights quick start]

You can group ratings into sentiment buckets directly in the UI without changing your data: [discover insights]

Negative: 1–2 ★
Neutral: 3 ★
Positive: 4–5 ★

This uses Mixpanel's custom buckets feature on the star_rating numeric property.

Filter to only show low-rated reviews (star_rating <= 2) and break down by platform or product_id to find problem areas.
Break down by user properties (e.g. subscription plan, country) to see if certain user segments leave lower ratings. [Insights quick start]

Goal	Report	Key Setting
Rating trend over time	Insights (Line)	Avg of star_rating
Rating distribution	Insights (Bar)	Breakdown by star_rating
Ratings by product	Insights (Bar)	Breakdown by product_id
Sentiment grouping	Insights	Custom buckets on star_rating
Low-rating deep dive	Insights	Filter star_rating ≤ 2

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai what is the ebst practice to display long text results in this kontext?

kapa.ai

APP

(edited)·

String properties are truncated at 255 characters. Any review text longer than that will be cut off. [property data types]
Free-text fields may contain unintentional PII (names, emails, etc.) — Mixpanel's privacy guidance explicitly warns: "Don't send unreviewed free-text." [data privacy guide]

Rather than sending raw free-text to Mixpanel, the knowledge sources recommend transforming long strings into structured, analysable properties: [import events]

"Parse the JSON and flatten it into properties... this is much more useful in analysis, as you can filter or breakdown by any key."

Applied to review text, this means categorising or tagging the text before sending it to Mixpanel:

Instead of...	Send this...
review_text: "Delivery was slow and packaging was damaged"	review_theme: ["Delivery", "Packaging"]
review_text: "The app crashes on checkout"	review_theme: "Bug", review_sentiment: "Negative"

This way you can filter, break down, and aggregate review themes meaningfully in Mixpanel reports.

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai how do I bets set up a report that shows the development of the average rating over time

kapa.ai

APP

(edited)·

Follow these steps in the Insights report: [Insights quick start]

1. Select your Event

Click "Select Metrics"
Choose Events → select Review Submitted

2. Change the Measurement

Click on the default "Unique Users" measurement
Select Aggregate Property → Average
Choose star_rating as the property

[measurements] 3. Set the Visualization

Choose a Line chart — this is the best chart type for showing a trend over time [Insights charts]

4. Set your Date Range

Choose a meaningful time window (e.g. last 30, 60, or 90 days)
Set the time interval (daily, weekly, or monthly) depending on your review volume

Enhancement	How
Smooth out spikes	Enable Rolling Average in Analysis Settings [analysis settings]
Compare to a previous period	Use Compare to Past (e.g. vs. last month) [time period comparison]
Break down by product/category	Add a Breakdown on product_id or review_category
Filter to a platform	Add a Filter on platform = iOS / Android

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai how can I prevent, that periods without active feedback are counted against the avearge

kapa.ai

APP

(edited)·

I'm sorry, but the knowledge sources do not explicitly describe a built-in Mixpanel feature that automatically excludes zero-activity periods from an Aggregate Property Average calculation.

The Average measurement in Insights calculates the average of a numeric property across all instances of the event. [aggregate property] This means:

If no Review Submitted events occurred on a given day, that day simply won't have a data point — it is not counted as a zero in the average calculation.
The risk is more about sparse data causing visual gaps or misleading spikes in the line chart, rather than zero-days dragging the average down.

Problem	Suggested Approach
Noisy spikes on low-volume days	Enable Rolling Average to smooth the trend [analysis settings]
Too many gaps in the line	Switch the time interval from Daily → Weekly or Monthly to aggregate more reviews per data point [date range]
Outlier periods skewing results	Use Compare to Past to visually identify anomalous periods [time period comparison]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
·
kapa.ai ratings can only be 1 through 5 - how can i prevent counting of 0 values on days that don't have feedback at all

kapa.ai

APP

(edited)·

When using Aggregate Property → Average on star_rating:

Mixpanel only calculates the average across actual instances of the event Review Submitted
Days where no reviews were submitted simply produce no data point — Mixpanel does not insert a 0 value into the average calculation [aggregate property]

So a rating of 0 should not appear unless your tracking code is explicitly sending star_rating = 0 as a property value.

This would most likely mean star_rating = 0 is being tracked in your data. In that case, you can exclude it with an inline filter:

1.
Click the … icon next to your Review Submitted event
2.
Add a filter: star_rating is greater than 0 (or more precisely: star_rating is between 1 and 5)

This ensures only valid ratings are included in the average calculation. [filters]

Scenario	Solution
Days with no reviews	No action needed — Mixpanel won't count them as 0
star_rating = 0 being tracked	Add an inline filter: star_rating > 0
Gaps/spikes from low-volume days	Use Rolling Average to smooth the trend [analysis settings]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

Tobias
·
(edited)·
kapa.ai how can i change the Type of properties from string to number, that are sent per integration (ie userflow)

kapa.ai

APP

(edited)·

You can handle this directly in Mixpanel using Typecasting, without needing to change anything in your integration (e.g. Userflow). [typecasting]

1.
Open any report (Insights, Funnels, or Retention)
2.
Add your property as a filter or breakdown
3.
Click the overflow menu (⋯) next to the property
4.
Select "Data Type"
5.
Change the type from String → Number

This forces Mixpanel to interpret the string value as a numeric type, enabling calculations like sum, average, median, and percentile. [typecasting]

Point	Detail
Report-level only	Typecasting applies only within that specific report view — it does not change how the data is stored
String must be numeric	The string value must be interpretable as a number (e.g. "4" → 4). If it cannot be parsed, the value becomes undefined [segmentation expressions]
Not persistent	You need to reapply the typecast each time you build a new report

If star_rating is being sent as a string (e.g. "4") by Userflow:

1.
Add star_rating as your breakdown or measurement property
2.
Typecast it to Number
3.
Then select Aggregate → Average to get your average rating trend

If you want the property to always behave as a number without manually typecasting each time, the best options are:

Fix it at the source — configure Userflow to send star_rating as a numeric value rather than a string, if the integration allows it
Use a Custom Property in Mixpanel to permanently define the transformation [custom properties]

ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

11 comments

11 comments

Analysing Customer Reviews (Star Rating + Free Text) in Mixpanel

1. Track a Single, Well-Attributed Event

2. Analyse in Insights

Average Star Rating Over Time

Distribution of Star Ratings

Ratings by Product or Category

3. Use Custom Buckets for Rating Groups

4. Filter & Segment for Deeper Analysis

Summary

Best Practices for Long Text (Free-Text Reviews) in Mixpanel

⚠️ Key Limitations to Be Aware Of

✅ Recommended Approach: Pre-Process Before Sending

Summary

Setting Up Average Star Rating Over Time in Insights

Step-by-Step Setup

Optional Enhancements

Tip: Rolling Average

Excluding Periods Without Feedback from the Average

What the Knowledge Sources Do Confirm

Practical Mitigations

Recommendation

Preventing Zero-Value Days from Affecting the Average

The Good News: Mixpanel Doesn't Inject Zeros

If You Are Seeing Zero Values

Summary

Changing Property Data Type from String to Number (Typecasting)

How to Typecast a Property in Reports

Important Caveats

For Your Star Rating Use Case

For a Permanent Fix

Analysing Customer Reviews (Star Rating + Free Text) in Mixpanel

1. Track a Single, Well-Attributed Event

2. Analyse in Insights

Average Star Rating Over Time

Distribution of Star Ratings

Ratings by Product or Category

3. Use Custom Buckets for Rating Groups

4. Filter & Segment for Deeper Analysis

Summary

Best Practices for Long Text (Free-Text Reviews) in Mixpanel

⚠️ Key Limitations to Be Aware Of

✅ Recommended Approach: Pre-Process Before Sending

Summary

Setting Up Average Star Rating Over Time in Insights

Step-by-Step Setup

Optional Enhancements

Tip: Rolling Average

Excluding Periods Without Feedback from the Average

What the Knowledge Sources Do Confirm

Practical Mitigations

Recommendation

Preventing Zero-Value Days from Affecting the Average

The Good News: Mixpanel Doesn't Inject Zeros

If You Are Seeing Zero Values

Summary

Changing Property Data Type from String to Number (Typecasting)

How to Typecast a Property in Reports

Important Caveats

For Your Star Rating Use Case

For a Permanent Fix