Targeting

[achafetz]

Targets need to be

1. divided up to determine weekly target and then create a cumulative target off that
2. merged into the main dataset

[achafetz]

Reopening with the question from Mavis;

With regards to the targets, will you be calculating a standard weekly target (e.g. annual target/52)? I was thinking that a calculation that includes the roll-over of unmet targets from prior weeks. Is the automation of such a calculation doable?

Currently, we calculate the weekly target as annual target divided 52 weeks.

If we moved to rolling, how would we handle "negative" targets (in visuals) where partner exceeded calculated weekly target? What if weeks are skipped during reporting?

[achafetz]

Maybe add additional visual on cumulative to track?

[achafetz]

Current Target process:

• Import from MER Structured Datasets, wpm_pulltargets()
• File is stored in the Outputs folder
• In running the wpm_combine() script, targets are incorporated via wpm_addtargets() by joining (left_join()) the above output (weekly targets) onto the weekly results.

This can be an issue because there may be missing weeks of reporting, which would then be missing the targets to join onto.

An alternative:

• Add in all the columns in the results file to the created targets file
• Filter for only the mechanism that are doing weekly targeting
• Duplicate each row 52 times, or the number of weeks into the FY (would need to make adjustments for TX_CURR, snapshot indicator)
• Conduct a full_join() with the results dataset

Since weekly targets are kind of meaningless, it may be best to use the cumulative results/targets graph (upper right above) to track progress