Blanks Auto-filling on Time Series, New Dataset Fields, and Email Matching on Login
This week's release highlight is a long-awaited feature: the ability to automatically fill missing data points as zeros on time series. This round of updates also brings some new fields to help filter on resolving PRs and resolved issues and some quality-of-life improvements on our login system.
Auto-filling on Time Series
Your time series charts may look slightly different since you last viewed them. That's because we now fill the blanks with zeros on time series, instead of "just" connecting the dots between values with underlying data. This produces sharper charts and better highlights periods of inactivity.
Let's take the example of a count of pull requests per month and assume the following data:
- July: 17 PRs
- August: 0 PRs
- September: 12 PRs
Previously, our time series would have connected July to September (2 existing data points), leaving an impression that there was some in-between count for August.
With the auto-filling feature, this time series will now display three data points, with the middle one being set to zero, therefore highlighting the lack of activity for August.
The auto-filling feature only fills the blanks between existing data points, not around them. So if you set your insight filters to retrieve data from the last 6 months but only have actual data for the last 4 months, the auto-filling feature will only fill in the blanks within those last 4 months.
If you still want to connect the dots between existing data points only, you can disable this auto-filling feature by editing your insights and going to the Customize Style tab:
New Fields on Datasets
The following new fields have been released to help you make your filtering even more powerful (click the fields to access their documentation).
Comments
Commits
Issues & Pull Requests
Reviews
Email Matching when Logging In with New Identities
When logging in to Keypup with a new identity (e.g. Google), we will now attempt to match this identity with an existing user by looking up identities with verified emails (e.g. GitHub) instead of creating a new Keypup account.
We were previously only matching identities that had been explicitly linked to your account upon logging. This new approach is a bit more flexible, especially when you do not remember which service you used to log in to Keypup.