Kerala DHS Facilities

[Rajesvari Parasa]

Hi,

I recently came across this dataset on govt health facilities put up by the Department of Health Services, Kerala. The file below has geocoded locations of these. It has details about sanctioned beds, phone number, and email-ID of each facility. (I generated address strings by appending the address elements in the columns and then used google geocoder API.)

Currently, the file attached here has locations of General Hospitals, CHCs, DHs, PHCs, THQHs and THs. There are some more types of facilities in the PDF e.g. Family Health Centres (FHCs), District TB centres, etc. Will clean those up and share those too.

The biggest caveat, however, is the fact that this is a geocoded dataset and will not be totally accurate. For example, there are instances where multiple facilities (even up to about 40 or so) have returned the same coordinate pairs. Details like street name and pincodes would have been helpful. Let me know if you have thoughts on how to improve this further. Thanks.

Hope this is of some use to some of you.

Best,

Rajesvari

[bob quest]

Hi Rajesvari

One solution around the pin code problem is to call up google search api and get the address with pin. I will update if I have any success.

Regards

Priyabrata Dash 

--
Datameet is a community of Data Science enthusiasts in India. Know more about us by visiting http://datameet.org
---
You received this message because you are subscribed to the Google Groups "datameet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to datameet+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/datameet/CAEKTdxp%3Dsz7tSAA4WKy19NR0fU7OK%2BVKoWd2vNzaqi%3Dpn4sZcg%40mail.gmail.com.

[Sunil Prabhakar]

Dear Rajesvari,

Great effort. Thank you very much. The geocoding may not be accurate due to several reasons. Same place name with different spatial data is one of the problems. The district data will help to minimise that error. Available pin code data is not found accurate again. The best method is to collect the complete postal address from the district offices and redo the geocoding with a much accurate postal address. Some crowdsourcing can also be tried to correct the data. I hope can explore more such ways.

Regards

Sunil

On Wed, Nov 4, 2020 at 12:35 PM Rajesvari Parasa <prkraj...@gmail.com> wrote:

--