Input Data Specification¶

Case Data Format¶

The primary input to the Acestor pipeline is the district-level daily dengue case data, stored in datasets/cases_district_daily.csv.

File Location¶

datasets/cases_district_daily.csv

File Format¶

The file is a comma-separated values (CSV) file with the following structure:

Column Specifications¶

Case Data Columns¶
Column Name	Data Type	Description
`state.name`	String	LGD Standard Name of the Indian state (e.g., “CHHATTISGARH”, “MAHARASHTRA”)
`district.name`	String	LGD Standard Name of the district within the state (e.g., “BALOD”, “DURG”)
`date`	String	Date of the observation in DD/MM/YYYY format (e.g., “01/05/2025”)
`samples_tested`	Integer	Number of samples tested for dengue on that date in the district
`case`	Integer	Number of confirmed dengue cases on that date in the district
`state.ID`	String	LGD Standard identifier for the state in the format state_lgd-code (e.g., “state_22”)
`region_id`	String	LGD Standard identifier for the district/region in the format district_lgd-code(e.g., “district_646”)

Data Characteristics¶

Temporal Coverage: Model performs better with more amount of data - Threshold calculation logic is of 2 types. One using historical data and other using past n weeks data. So data must either be recent upto current day, or there must be data of previous years for the model to work accurately.
Spatial Granularity: District level
Geographic Scope: Indian states and their districts
Primary Metrics:
- Number of samples tested per day
- Number of confirmed dengue cases per day

Example Records¶

state.name,district.name,date,samples_tested,case,state.ID,region_id
CHHATTISGARH,BALOD,01/05/2025,1,0,state_22,district_646
CHHATTISGARH,BALODABAZAR-BHATAPARA,01/05/2025,3,0,state_22,district_644
CHHATTISGARH,DURG,01/05/2025,22,0,state_22,district_378

Data Quality Notes¶

Case Values: Case counts are non-negative integers
Date Format: All dates follow the DD/MM/YYYY format
Identifiers: Both state.ID and region_id are based on LGD Standard.