Skip to content

Home

Jump to bottom
Alexandersax edited this page Aug 30, 2012 · 15 revisions

#Woohoo! You made it this far!

###Before we do anything else, let's get you set up with GitHub for Windows.

Click here once you installed.

This wiki assumes you have a working knowledge of STATA and .do files as well as an idea of what packages are. If you're unsure about packages or anything else, use the help command in stata or click the magic link.

I'm sorry if you clicked that link, that link was rude. Can forgive each other and move on? Awesome! Let me start over.

###Welcome to the wiki and repo for RTI's STATA code. This wiki will give you a rough overview of how it's done.

This wiki assumes you have a working knowledge of STATA and .do files as well as an idea of what packages are. If you're unsure about packages or anything else, use the help command in stata or refer to google.

Let's start off with how we process incoming data. It'll go through a few stages. Glance at the chart below, and then read the description.

workflow

  1. Raw data comes in from the field, usually in an Excel file. We import that into Stata and move on to #2

  2. Cleaning data (demographic variables, EGRMAclean data, weighting) in a 2.do file

  3. Cleaned data is ready to be analyzed or published

  4. Analyzing data, sending to client, getting new requests, reanalyzing - all this gets done in a 4.do file

  5. The information then gets used in reports that inform policymakers who then set policies which improve schools.

That's all there is to it. Steps 2 and 4 used to eat up entire weeks, but now require far less time because we've automated (much of) those tasks. The turnaround time for data has shrunk from 2-4 months to under a week. This is the way it should be! There's no reason for data to spend time idle on a server.

##Data should be easy, for everyone

Although we've made vast improvements in or data cleaning methodology, there's still work to be done! The code must be maintained and improved to streamline our process. As EGRA scales, so too much our code - we hope to offer more analysis features and perhaps (cross your fingers) automated weighting.

###So how do we take care of our code? Taking care of our code usually falls under two categories

For bottom-up information, beginning with how the packages in STATA work, click here.

Clone this wiki locally