Data Analysis

Almost every project we undertake involves data analysis of some kind. In most cases the preliminary analysis is done in Excel by sorting and filtering the data, or by creating pivot tables to discover the correlations between different data fields. 

We also use a variety of analytical tools if specialised analysis is needed. These tools can employ statistical analysis (EViews, Stata) or geospatial visualisation (Google Fusion Tables, QGIS, Tableau, CARTO). Our scope of work will include a determination of which tools are suitable for the job at hand.

Case Studies

Matric results analysis 

Every year in January the results of the national matriculation exams are released. The typical stories the media tell are either the story of an aggregated result at the provincial or national level, focusing on an overall pass rate, or stories of individual excellence, normally about matriculants taking the Independent Examination Board (IEB) exams.

We feel that these stories miss the important trends and developments taking place at a dissagregated level. Since 2015 we’ve focused on analysing matric results at the school and subject level.

The Department of Basic Education (DBE) provides detailed results by school and by subject. We created a database of matric results that now covers eight years of data from 2012 to 2019.

Our analysis is able to pinpoint how many distinctions were achieved in a particular subject in a particular school, how the average mark per subject has changed in a school or a region, and so on.

We partnered with the Mail&Guardian in 2017 to write a series of articles on matric achievements. We also worked with ORT SA, a non-profit organisation which works with provincial education departments.

ORT SA has been working with high schools in the Alexandra township in Johannesburg to improve education outcomes. Their work began with the 2007 school cohort which wrote matric in 2019. 

We were able to show that matric results in mathematics in the five schools where ORT SA is involved were better than in other high schools around Alexandra. Since all the schools in the township cater to learners with similar demographic and developmental backgrounds we were able to conclude that the ORT SA interventions were the distinguishing factor in achieving better education outcomes. 

Analysis of basic food prices

There are about 15 major fresh produce markets in South Africa. Over 100 different fruits, vegetables, herbs and other food commodities are sold daily at these markets. 

We built a data-scraping tool to capture data in the three largest markets: Johannesburg, Cape Town, and Durban. The tool scans data from the markets’ websites and copies it to a database. 

We have over four years of daily prices and quantities sold for dozens of food commodities. We analysed this time-series data and built an economic / statistical model from our results.

We estimated how the price of popular food items changes with a change in market supply and we measured the seasonal effect on prices for some of the rarer items.