bellingcat/instagram-location-search
1
0
Fork 0
mirror of https://github.com/bellingcat/instagram-location-search.git synced 2026-06-08 02:28:29 +03:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #1 from nemec/main

Browse Source 
Add a --date filter and some other quality-of-life improvements to the generated data
This commit is contained in:
Logan Williams
2020-12-18 08:42:25 +01:00
committed by GitHub
parent fea4c07659 bd51e92115
commit 8ca974319b
2 changed files with 52 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
README.md
View File

@@ -2,7 +2,7 @@
## Prerequisites
This Python application requires `requests` and `numpy` to be properly installed. This can be done with `pip3 install requests numpy`.
This Python application requires `requests`, `numpy`, and `pandas` to be properly installed. This can be done with `pip3 install requests numpy pandas`.
## Example usage
@@ -12,6 +12,23 @@ The following command will search for Instagram locations nearby the coordinates
Note that this requires an Instagram session ID in order to work! See below for how to obtain one from your account.
## Example usage with date
The following command will search for Instagram locations near Seattle's "Capitol Hill Autonomous Zone" during the George Floyd
protests in early June, 2020. Not all location pages in the area will have posts relevant to the Zone, but some do. Open the
resulting `map.html` file in your browser to view locations.
```python3 instagram-locations.py --session "<session-id-token>" --lat 47.6164311 --lng -122.3203952 --map map.html --date 2020-06-09```
When using the `--date` argument, links to Instagram location pages will be filtered to show posts created on this date or earlier.
Instagram will usually first show a 3x3 grid of "Top Images and Videos" that are more recent, however once you scroll past that
there is a section labeled "Most recent" which will show the posts sorted by date (if any).
These links are only used together with the `--csv` and `--map` arguments, they aren't included in `--json` or `--geojson`.
Note: Instagram treats these dates as "UTC", which is a timezone near Great Britain. If your target location is far from this zone,
it's worth adding a couple of days to your filter to make sure you capture all relevant posts. Also, this only specifies the
*maximum* post date that can be displayed. If nothing was posted that day at that location, it will show older posts (sometimes
even multiple years older).
### Other output formats
Using the `--json <output-location>` command line argument, the list can be saved as a JSON file, almost identical to the raw API response.

38
instagram-locations.py
View File

@@ -4,6 +4,8 @@ import pandas as pd
import argparse
import json
from string import Template
from datetime import datetime, timezone
import sys
# gets instagram "locations" around a particular lat/lng using internal API
# (requires session cookie for authentication)
@@ -11,7 +13,6 @@ def get_instagram_locations(lat, lng, cookie):
locs = requests.get("https://www.instagram.com/location_search/?latitude=" + str(lat) + "&longitude=" + str(lng) + "&__a=1", headers={
'Cookie': cookie
}).json()
return locs['venues']
@@ -55,6 +56,23 @@ def make_geojson(locs):
return {"type": "FeatureCollection", "features": features}
def encode_date(date_str: str):
'''Convert date into Instagram "snowflake" ID'''
try:
date = datetime.strptime(date_str, '%Y-%m-%d')
except ValueError:
try:
date = datetime.strptime(date_str, '%Y-%m-%d')
except ValueError:
print('Unable to parse date. Please use format "yyyy-mm-dd".', file=sys.stderr)
sys.exit(1)
date = date.replace(hour=23, minute=59, second=59, tzinfo=timezone.utc)
date_ts = int(date.timestamp()) * 1000 # milliseconds
insta_epoch = date_ts - 1314220021300
max_id_num = insta_epoch << 23
return str(max_id_num)
html_template = '''<html>
<head>
<title>Instagram location visualizations</title>
@@ -84,6 +102,10 @@ html_template = '''<html>
width: 100%;
height: 600px;
}
img.selected-location {
filter: hue-rotate(120deg);
z-index: 999 !important;
}
</style>
</head>
<body>
@@ -102,12 +124,15 @@ html_template = '''<html>
}).addTo(map);
function onEachFeature(feature, layer) {
layer.bindPopup(`<a href="https://www.instagram.com/explore/locations/` + feature.properties.external_id + `">` + feature.properties.name + `</a><br />` + feature.properties.address );
layer.bindPopup(`<a href="https://www.instagram.com/explore/locations/` + feature.properties.external_id + `$date_var" target="_blank">` + feature.properties.name + `</a><br />` + feature.properties.address );
}
L.geoJSON(locs, {
onEachFeature: onEachFeature
}).addTo(map);
var centerMarker = L.marker([$lat, $lng]).addTo(map);
centerMarker._icon.classList.add('selected-location');
</script>
</body>
</html>'''
@@ -121,11 +146,16 @@ def main():
parser.add_argument("--csv", action="store", dest="csv")
parser.add_argument("--lat", action="store", dest="lat")
parser.add_argument("--lng", action="store", dest="lng")
parser.add_argument("--date", action="store", dest="date")
args = parser.parse_args()
cookie = 'sessionid=' + args.session
date_var = ''
if args.date is not None:
date_var = '?max_id=' + encode_date(args.date)
locations = get_fuzzy_locations(float(args.lat), float(args.lng), cookie)
if (args.output):
@@ -136,7 +166,7 @@ def main():
if (args.map):
s = Template(html_template)
viz = s.substitute(lat=args.lat, lng=args.lng, locs=json.dumps(make_geojson(locations)))
viz = s.substitute(lat=args.lat, lng=args.lng, locs=json.dumps(make_geojson(locations)), date_var=date_var)
f = open(args.map, 'w')
f.write(viz)
@@ -144,7 +174,7 @@ def main():
if (args.csv):
df = pd.DataFrame(locations)
df['url'] = df['external_id'].apply(lambda v: 'https://www.instagram.com/explore/locations/' + str(v))
df['url'] = df['external_id'].apply(lambda v: 'https://www.instagram.com/explore/locations/' + str(v) + date_var)
df.to_csv(args.csv)
if __name__ == "__main__":