For many LiveRampers, hackweek is one of their favorite parts of working here. It’s a time when you can work on whatever you think will bring the most value to the company, whether it is fixing a process that’s been making you miserable, cleaning up some old code that’s dragging down the system, or exploring an off-the-wall new technology. In fact LiveRamp actually started off as a hackweek project. If you enjoy having the autonomy to work on anything you want, we’re always hiring. Below are some of the projects from hackweek XLIV.
Swagger is a really cool tool for building and working with APIs and is used extensively through our codebase. From a swagger config file it generates clean, pretty API docs and it can even be used for code generation. One big problem though is that swagger config files are difficult to generate and work with; minor changes can cause the whole file to not compile and lead to a slew of errors. Because of lack of familiarity with swagger, many engineers will copy over an existing swagger config file and modify it to fit their API – but the workflow of slogging through the resulting errors is very inefficient.
Swagoo presents an alternative workflow: the swagger GUI (haha, took us a sec to get the name). The GUI walks the user through the process of building a swagger config file and handles the pitfalls that would cause the file to not compile. The GUI is still a work in progress, but we’re excited to see more features get added and be used in our workflows!
Contrary to what you might think from the name, this hackweek project is actually intended to help increase scope creep – for permissioning scopes, that is. At LiveRamp, our permissioning system is based on personas and scopes. A persona is a general type of customer that a client fits into, while scopes are individual pieces of functionality that can be either shown or hidden. Each persona has associated scopes that it is granted access to, and each client is categorized as one or more personas.
Segment Correlation Explorer
At LiveRamp we see massive amounts of data flow through our system, but our core product is sharing that data across different anonymous identifiers (think cookies, hashed email addresses, mobile IDs) rather than performing analysis on the data. This group thought it would be cool to investigate what correlations there are between different segments (think characteristics about a person) given that we have access to such a massive data set.
They built a visualization tool where you can input two or more features you want to compare and it returns a graph with nodes representing attributes and weighted edges showing the strengths of correlations. This lead to some interesting results; for instance the correlations between political party and sports below. Congratulations to Segment Correlation Explorer for winning the judges choice award!
This was only a sampling of the projects that came out of this hackweek, and they happen once per quarter so the next one is less than three months away! Excited to see what new innovation will come then.