search

Science-for-Nature-and-People / soc-twitter

SNAPP - Soil Organic Carbon Twitter data
1 stars 4 forks source link

Map all tweets #2

Open swood-ecology opened 5 years ago

swood-ecology commented 5 years ago

Map the location of all tweets related to different categories. First focus just on "soil health" terms. Compare that to "rangeland health", "forest health", "soil quality", etc. If soil quality and soil health overlap strongly geographically, could group into one map.

brunj7 commented 5 years ago

@grantnolasco already added the country code to the data set and developed a script to do few maps.

@grantnolasco can please push the script to this repo.

swood-ecology commented 5 years ago

@brunj7 did this ever go anywhere? this topic seems to be a priority for folks in the group.

swood-ecology commented 5 years ago

@brunj7 Are there lat-long points in the data set? Can we replicate Figure 6 in this paper: https://www.academia.edu/20543286/Analyzing_the_language_of_food_on_social_media

swood-ecology commented 5 years ago

Also, it would be good to have the map color the points by whether they're part of the original database, or the regularly updating database.

swood-ecology commented 5 years ago

From @kylemonper: not collecting lat/long for regularly updating tweets and very few tweets from original data have this.

kylemonper commented 5 years ago

was able to get about 5,500 locations by using the google api to geocode the location info provided by some of the tweets:

map_prov

map_query

This is still rough, id like to explore different base maps/remove the country names, mess with point sizes, etc. Perhaps even zoom into individual countries to look at spatial distribution of query terms

swood-ecology commented 5 years ago

Thats fantastic! That’s exactly what I was thinking. Is that all in a separate script on GitHub?

On Oct 17, 2019, at 9:44 PM, kylemonper notifications@github.com wrote:

was able to get about 5,500 locations by using the google api to geocode the location info provided by some of the tweets:

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

kylemonper commented 5 years ago

its currently on the influencer branch within the map folder -- I'm going to do a big merge to the master tomorrow

swood-ecology commented 5 years ago

I was thinking a bit more about this and this map would make a really cool starting point for looking at regional differences in language co-occurring with the major search terms (soil health vs regenerative). It would just be a subset of the overall dataset but I think it would be really interesting to look at that for North America vs Latin America vs Europe vs Africa vs Australia.

On Oct 17, 2019, at 11:09 PM, kylemonper notifications@github.com wrote:

its currently on the influencer branch within the map folder -- I'm going to do a big merge to the master tomorrow

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

brunj7 commented 5 years ago

maybe also switch to a les colorful background, maybe simply country borders