Full data extraction and cleaning
- lisamatay
- Aug 6, 2024
- 2 min read
Data cleaning process
This week started off on a very high note following the encouraging and thought-provoking discussions at the Ministry of Health. As a data analyst and researcher, it is extremely important to be knowledgeable about the context and understanding of the landscape as you work on data analysis. With that context in mind, I was able to proceed efficiently in data cleaning – another very crucial but painstaking task requiring meticulous attention to detail.
This step has to be thoroughly carried out as it lays the foundation for future analyses therefore adequate time should be spent here to ensure this. This week and the next will be spent on my laptop extracting data from the DHIS2 system for all facilities in Ethiopia. The country has about 40,000 public health facilities. Colleagues at the office were very helpful in quickly reconciling any discrepancies I observed while data cleaning. In particular, Ethiopia has 9 regions and 3 administrative cities. Understanding the administrative hierarchy in each region is crucial to data cleaning as those differences often show up in the way the data is structured on the DHIS2 platform. The goal is to create a comprehensive dataset of all public health facilities in Ethiopia sorted by type of facility (health post, health center, hospitals) and indicators for in-facility delivery, skilled birth attendance, post-natal and antenatal visits. This dataset will then be used to generate useful insights for the Ministry of Health in regards to maternal and child mortality.
Cultural experiences
I was very excited to attend a couple of cultural events including an ethio-jazz event that uses a combination of "Western" and traditional Ethiopian instruments with important historical and cultural significance.
One example is during the battle of adwa against the Italians, they were used to mobilize people as a call to action and to generate morale in the resistance movement.
Comentarios