It’s people closest to the work, within school communities - teachers, young people, families - who will be able to build the next generation of school data systems that they need.
We need more than just systems for counting numbers - we need ways to more deeply connect and tell our stories to tackle what matters for our students.
Learn more at studentinsights.org.
Check out the demo site with different roles:
All accounts use the password:
We'd love your help! Take a look at CONTRIBUTING.md for more information on ways educators, developers and others can get involved and contribute directly to the project. You can also learn how to join our online chat channel and submit pull requests and join us in person at our weekly hack night with Code for Boston, in Kendall Square, Cambridge.
This is a Ruby on Rails app that uses a PostgreSQL database as the primary data store, and relies on React for much of the UI code.
You'll need Ruby, Postgres and yarn. See our local installation on OSX or Linux guide.
$ bundle install $ yarn install
bundle exec rake db:create db:migrate db:seed
This will create demo students with fake student information. See the demo site above for the set of educators you can use (or look at
If you are willing to run a longer (~10 minute) task that will generate ~600 students to more closely approximate one of our pilot schools, set
ENV["MORE_DEMO_STUDENTS"] = 'true' before running the seed task.
Once you've created the data, start the app by running
command+c like normal, and look at
package.json if you want to run them in individual terminals.
For Ruby code, to lint and run the tests do:
For Jest, run the tests continually in watch mode with:
rubocop yarn lint
If you miss something, tests will run on any pull request you submit, and after merging to master as well.
If you use Sublime Text Editor, we include the
studentinsights.sublime-project file to standardize minor things like tabs vs. spaces, indentation, and generally make it easier to work with our folder structure. Go to
Project --> Open Project and select that file to load it. Sublime remembers which project you were last in, so you only need to do this once. (Here's some background info on how projects work in Sublime).
We also recommend Sublime Package Control and these packages Babel, Sublime Linter and SublimeLinter-contrib-eslint. These will give you nice syntax highlighting and show you linter errors right in Sublime!
Users use IE11, so if you're trying to manually test locally or the production site, you should too! If you have a Mac or Linux box, you can use free VMs designed for just this purpose and run them on VirtualBox: https://developer.microsoft.com/en-us/microsoft-edge/tools/vms/.
yarn storybook will start a storybook server on port 6006. You can use this to create "stories" iterate on UI features or components in particular states, separate from how they exist within the product. To add new stories, write a new
.story.js file and update
# prompt GIT_PS1_SHOWCOLORHINTS=1 GIT_PS1_SHOWDIRTYSTATE=1 GIT_PS1_SHOWUNTRACKEDFILES=1 GIT_PS1_STATESEPARATOR=' --!-- ' source ~/.git-prompt.sh PS1='\h:\[\033[1;34m\]\W \[\033[0;33m\][$(__git_ps1 "%s")]\[\033[0;37m\] \$ '
You can find a copy of
git-prompt.sh in source.
The project is a Rails app with a Postgres database. There are background tasks that import data from district SIS systems, vendor systems like STAR, and district data sources within Google Drive. This enables rapid iteration and experimentation on new product features with minimal risk to these existing production systems. Student Insights is the system of record for some unique data (eg., notes from clinical meetings and information about targeted interventions that students are receiving). Authentication is typically handled by the district's LDAP service.
District project leads have access administrative tools for adjusting educator permissions and roles, for doing data quality checks, reviewing health of import tasks, and exporting data.
Understanding the human pieces of how data flows through the system, which is critical for understanding accuracy and representativeness, is quite complex and varies within and across districts. This is important to keep in mind when making product and engineering decisions. With that caveat, here's an abstracted engineering-centric view on the pieces of the system and kinds of data flow (from "Student Insights - How it works" internal doc):
All districts share the same codebase, but are deployed in isolated instances.
There are many differences within and across districts. Some of this is driven by districts focusing on different things or being at different places in adoption, and some is different by inherent differences. When trying to learn about or understand how Student Insights handles these differences, a good triage process is:
PerDistrict.jsin source code
EducatorLabelwhich describes different feature switches, which are often related to important differences across or within districts
For educator-facing document, see also:
We use GitHub Actions to run a set of tests on each pull request and merge to master. See the
actions.yml file for more, and view test runs it GitHub at https://github.com/studentinsights/studentinsights/actions. There's no automatic deployment set up, so nothing will go into production until someone takes action to deploy it.
Access to production environments is tightly controlled. If you work in a particular district or have been granted access to production systems, talk to someone on the team. See CONTRIBUTING.md.
Quotaguard Static, a Heroku add-on, provides static IP addresses when needed to connect to firewalled district servers. The
QUOTAGUARDSTATIC_MASK environment variable is a subnet mask for routing only certain outbound requests through those proxy servers. Read Quotaguard Static's documentation for more information.. See the Procfile for how this is enabled; the socksify wrapper makes routing through the proxy transparent to Ruby application code.
To create an entirely new Student Insights deployment, ask someone for the "New Student Insights deployment" doc.
Here are some notes on maintaining, troubleshooting and performance. Talk to someone on the team for access to more tools and docs.
Look in the Heroku metrics panel.
You can use heroku-pg-extras to get helpful diagnostic information about slow queries, index usage, and table scans.
Heroku Postgres supports a maintenance window for standard database operations that require minutes of downtime. See