search

jeremysze / LPIS

Repository of code used to analyze LPIs in NYC
0 stars 0 forks source link

Results from spatial model #11

Open jeremysze opened 5 years ago

jeremysze commented 5 years ago

It seems the results from the spatial analysis is quite different from the non-spatial analysis. It could be because this is just Manhattan.

In the spatial lag model:

In the spatial error model:

Running the spatial lag and error model now.

jeremysze commented 5 years ago

https://github.com/jeremysze/LPIS/blob/master/stata_spatial_regression_qt.ipynb

I decided to do some check and ran the xtreg and xtpoisson on collision counts and number of person injured. It showed that LPIs is no longer significant in reducing collision counts in both xtreg and xtpoisson. I'm no sure how to make sense of this, because in the spxtreg model, LPIs remained significant in reducing collision counts.

For number of persons injured, LPIs remains significant for both models.

Maybe for the abstract, I will change focus to number of persons injured.

jhconning commented 5 years ago

That sounds good...

On Thu, Feb 14, 2019 at 12:23 PM Jeremy Sze notifications@github.com wrote:

https://github.com/jeremysze/LPIS/blob/master/stata_spatial_regression_qt.ipynb

I decided to do some check and ran the xtreg and xtpoisson on collision counts and number of person injured. It showed that LPIs is no longer significant in reducing collision counts in both xtreg and xtpoisson. I'm no sure how to make sense of this, because in the spxtreg model, LPIs remained significant in reducing collision counts.

For number of persons injured, LPIs remains significant for both models.

Maybe for the abstract, I will change focus to number of persons injured.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/jeremysze/LPIS/issues/11#issuecomment-463714397, or mute the thread https://github.com/notifications/unsubscribe-auth/AE0JH48gKbFyyhawva_rSwG1Aio-z3fOks5vNZuTgaJpZM4a51fW .

jeremysze commented 5 years ago

I also sent an email asking for some advise on the abstract sections.

jhconning commented 5 years ago

So I guess my instinct on the 'two abstracts' question is why not just send one abstract to that conference.. I don't see you doubling or greatly increasing chances by submitting two and as you suggested, you're running the risk of being excluded... Since you've applied to some other conferences to me that seems the better strategy... but whatever you think

On Fri, Feb 15, 2019 at 12:05 PM Jeremy Sze notifications@github.com wrote:

I also sent an email asking for some advise on the abstract sections.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/jeremysze/LPIS/issues/11#issuecomment-464124818, or mute the thread https://github.com/notifications/unsubscribe-auth/AE0JH1Z8OHlBH_73SEjD49neLZ8hdoFPks5vNujYgaJpZM4a51fW .

jeremysze commented 5 years ago

http://rstudio-pubs-static.s3.amazonaws.com/5027_52298866e7924b18b54e5c9a0a21b450.html

Saving this here for future reference