Significant drop in death count

[jhtep]

The death count has drop on 12/07
For example zip code Brooklyn 11234
Check from the New York Times web site and tracked back to this repo.
File: https://github.com/nychealth/coronavirus-data/blob/master/totals/data-by-modzcta.csv

Date	Cases	Per 100,000	Deaths	Per 100,000
12/07	3,295	3,457	187	196
11/30	3,144	3,299	198	208
11/19	2,975	3,122	197	207
11/04	2,828	2,967	196	206
09/17	2,526	2,650	195	205
...
07/23	2,353	2,469	195	205
..
07/08	2,313	2,427	192	201
...
07/03	2,291	2,404	191	200

[M. Montesano]

Thanks for the question. We do constant data cleaning to ensure data quality; all data are preliminary and subject to change. As mentioned in the documentation, the 12/7 update reflects a new geocoding process that causes some zip-level data to change slightly.

[jhtep]

Even with a slight change to zip-level data, shouldn't the overall total be same or increase?

Zipcode 11234 drop by 11 deaths, shouldn't the 11 deaths show up in another zip code?

I see a city wide drop in death stats on 12-07.

I extracted the historical data for totals/data-by-modzcta.csv and
I see a significant drop in deaths totals on 12-07
that is NOT balanced out across zip codes.

Here's what I see -

-- Total deaths in NYC during period 11-30 to 12-09:
19,550 New York City Deaths to date 2020-11-30
19,555 New York City Deaths to date 2020-12-01
19,625 New York City Deaths to date 2020-12-06 ********
19,288 New York City Deaths to date 2020-12-07
19,306 New York City Deaths to date 2020-12-09