Hi all, we are using program indicators to calculate some reports in dhis2, and we also need to compare the reports with reports with the reports (same report) from another system using the same data but not in dhis2. Some program indicators give different results. I tried using event report to export data and calculate in excel, the result also different from program indicator.
Is there anyway i can get the list of teis/events that counted in program indicator? Just the uid is enough.
In general you should use Event reports, where make a list of all events and attributes which was involved in calculation of the indicators for the given period and given OU(s). Generate a list and export it as CSV, not Excel. The reason is when you export as CSV it will be easy to separate all empty cells from the cells with values. You can check it with COUNTA function, which will give you more precise number of empty fields.
Then make manual calculations on that CSV file and compare indicators values. if there is any discrepancy you will understand why it is happening by filtering particular column on CSV file. Also during the analysis I would recommend you to use COUNTIF function to calculate particular values for the fields with options.
The rest is about how you master EXCEL in order to calculate and compare your data against DHIS indicators.
Also you can produce a pivot table from CSV and compare data.
I did download and use excel to calculate the indicator. The problem is the excel result is also different from PI result. I tried to change the boundary and filter but still it didnt help. The indicator itself is complicated and I dont even sure if the PI is calculating it the same way i did in excel.
@lehuyen so you should first make “right” calculations on EXCEL and check your filters and expressions.
Regarding boundaries if it’s not complicated one, the standard boundaries should be OK.
There might be also something connected to the data, where maybe field contains SPACE or any other data. It may reflect different results.
Otherwise we should look at the initial formula, your PI and your data all together.