Kolinx25/Ebola_Analysis_with_python

This project analyzes 200 confirmed Ebola cases from the early months of the 2014 West Africa outbreak in Sierra Leone. It combines epidemiological analysis with machine learning to answer two questions: how fast was the outbreak growing, and what factors predicted whether a case was detected quickly or late? # Ebola Outbreak Analysis — Sierra Leone, 2014 This project analyzes 200 confirmed Ebola cases from the early months of the 2014 West Africa outbreak in Sierra Leone. It combines epidemiological analysis with machine learning to answer two questions: how fast was the outbreak growing, and what factors predicted whether a case was detected quickly or late? I originally worked on a version of this in R. This is the Python rebuild — extended with R0 estimation and a Random Forest classifier. --- ## What's in here --- ## Dataset Each row is a confirmed Ebola case with the following fields: | Column | Description | |---|---| | id | Case identifier | | age | Patient age in years | | sex | M / F | | status | All confirmed in this subset | | date_of_onset | Date symptoms began | | date_of_sample | Date sample was collected | | district | District where case was recorded | Source: Sierra Leone Ministry of Health / WHO 2014 outbreak records. --- ## What the analysis covers **1. Demographics** The median case age was 35 years. Women accounted for 57% of cases (114 female, 86 male), which aligns with documented patterns of female caregivers having higher exposure risk during the West Africa outbreak. **2. Geographic distribution** Kailahun district accounted for 155 of the 200 cases — 77.5% of the entire dataset. Kenema was a distant second at 34. This reflects the outbreak's origin point near the Guinea border and the initial concentration of cases before spread to other districts. **3. Temporal trend** Cases were recorded between May 18 and June 30, 2014. The single highest-incidence day was June 10, with 20 new onset cases. A 7-day rolling average is plotted alongside daily incidence to smooth day-to-day noise. **4. R0 estimation** Using the exponential growth rate method on cumulative cases during the first 30 days, and applying Ebola's published serial interval of 15 days (Camacho et al., 2014): R0 = 2.79 means each infected person was generating roughly 3 se …