Progetto parte del corso Passion in Action Social Media in Emergency Rapid Mapping.
Indicazioni:
-
PP1
- import annotated table (Sandy Dataset)
- compute Vincent distance for two tweets, given their ids
- plot a table showing the total number of: not annotated, n.a., [], annotated with one location, annotated with two locations or more visualize them on a map PP2
- import the full Sandy dataset .json file (benchmark_ny_annotated.withcopyright.json)
- find the annotated locations in the and put them in a separate table. To do so, use the ""mention"" tag in the .json file. For eg: ""mentions"": [{""indices"": [37, 44], ""class"": ""Location"", ""subclass"": ""admin"", ""name"": ""new york""}] ) dice in sostanza (sull'esempio dato) la parola New York e' menzionata nel testo del tweet dal carattere 37 al 44
- find locations with the ""name"" string of the full dataset inside New York City with Nominatim
- compute the Vincent distances between locations of our annotated set and the full dataset and store them (if more than one location compute all combinations) PP3
- (optional) try some analysis