Correct the identified data point

wengxyu1030 / UHC-Time-Trend

This repository is to build time series for UHC indicators by country, year, survey.

0 stars 0 forks source link

Correct the identified data point #28

Open wengxyu1030 opened 2 years ago

wengxyu1030 commented 2 years ago

[ ] The list of problematic data points identified needs correction: (a spread sheet link to be attached to track the progress)
[ ] The list of problematic data points identified needs to be dropped: to be delivered as commend to the team lead.

Tracking sheet: Task Tracker: Indicator Review

Meeting Note: https://docs.google.com/document/d/15S5_M2lyUuguEvbKP9LqAFa4AeQJyI98-v5z-QjLkqA/edit

robin-wang commented 2 years ago

Copied from working issue for RW and AW. Task Tracker (Archived): https://docs.google.com/spreadsheets/d/1LTD05fPVJ-HItncaLC5orBk2cmRG2QDuOuvrkw_V4tw/edit#gid=1837304965

robin-wang commented 2 years ago

RW: prioritise anc, delivery, treatment AW: to add a spreadsheet

robin-wang commented 2 years ago

With Sven's proposed update, we are going back to all Central Asian Countries to make sure their maternity homes/hospitals are treated as hospitals.

robin-wang commented 2 years ago

@wengxyu1030 Hiya Aline, issues I'd like to note regarding the corrections in dashboard you are making.

I saw you started removing the data points Sven identified and we coloured in purple. Could you please make two exceptions?

MICS Guinea2018, we want to bring that back to Guinea-Bissau
MICS Congodr2014, we want to bring that back to CongoRep

My preliminary exmination of the potential corrections reveals a largely reassuring distribution when these two surveys are put into where they belong, but a few vars do show up less than perfectly. It might help to see them in the distribution.

Many thanks!

wengxyu1030 commented 2 years ago

Haven't updated the dashboard yet, but thanks for your note, will keep this in mind when an update is needed.

@wengxyu1030 Hiya Aline, issues I'd like to note regarding the corrections in dashboard you are making.

I saw you started removing the data points Sven identified and we coloured in purple. Could you please make two exceptions?

MICS Guinea2018, we want to bring that back to Guinea-Bissau

MICS Congodr2014, we want to bring that back to CongoRep

My preliminary exmination of the potential corrections reveals a largely reassuring distribution when these two surveys are put into where they belong, but a few vars do show up less than perfectly. It might help to see them in the distribution.

Many thanks!

wengxyu1030 commented 2 years ago

About MICS Guinea2018, the label of the raw data folder is not consistent which makes the code read the wrong string. Changing folder name to "MICS6-GuineaBissau2018". @robin-wang Please do so on your end too.

wengxyu1030 commented 2 years ago

@robin-wang Corrected the case for congodr2014 to congorep2014 in https://github.com/Data-Whale-DC/WB-HEFPI-MICS5/commit/967721af6c9e3afb7fb328d6a96c160207061d97

robin-wang commented 2 years ago

About MICS Guinea2018, the label of the raw data folder is not consistent which makes the code read the wrong string. Changing folder name to "MICS6-GuineaBissau2018". @robin-wang Please do so on your end too.

I looked at the output and raw data folders on my end, seems all were named Guinea-Bissau2018. Shall we keep names standardised to "GuineaBissau2018"?

robin-wang commented 2 years ago

@wengxyu1030 I am adding a list of variables to be dropped, based on the task tracker sheet. When you feel better, we could chat about the best approach to integrate removing indicator values, and other steps. See email.