The new geolocation analytics
It’s our great pleasure to introduce the most advanced geographical based social media analytics. At first glance it looks just like the old heatmap (the one you were used to in uberVU and the one most social media packages were offering). But trust us, it’s not your grandmother’s map. We introduced so many goodies that the our old map is like an Egyptian scroll compared to an iPad.
Zoom, Hover & Click
Our first frustration with our old map was its lack of interactivity. You had a world heatmap, all colored and nice but there was nothing more you could do really. And that alone was never good enough for us. We now have a fully interactive map. You can zoom, if you hover over a region it gets highlighted and it shows you the most important metrics. Want to go deeper? Click on a region to highlight the sub-regions.
Another thing that you’ll notice is that the whole page reacts to the changes on the map. Zoom on a country and the metrics below adjust to reflect the selection you made.
States & Cities
But the most important part is you can zoom on a state level and then further down to a city level. Yes. You can see how many mentions “Android” got in San Francisco, and how many in New York City.
On the world map, click on a country to zoom and reveal sub-regions. For a big country like US or Brazil you’ll be able to drill down to regions first. This will show the most popular states or regions. Hovering on a region will reveal the main metrics, while clicking will reveal the top cities. Hover over a city to get more info, or just look below the map for more info on the region: top cities, top languages.
How does it work?
Geotagging was always hard to do in social media. While there are a lot of people who fill in their location details in their profiles on Twitter, Facebook or Youtube, the data can be hard to interpret and/or be unavailable to companies like us. Twitter and Facebook introduced location based status messages, but very few messages have this type of meta information. So how do you do it then?
We decided to create a smart agent that considers all the potential inputs and makes an “informed” decision based on the information it has. Some of the things we look at:
- geolocation mention: if Mark tweeted and geotagged his tweet we’ll know precisely where he’s based. We’ll know the city, state and country. That’s not the case with most mentions though.
- profile description: Paulo said he is from Rome. If this is the only information we have we’ll assume he is in Italy.
- language: If Paulo said he is from Naples, that’s a little confusing. There are more than a dozen cities called Naples in the world. If he speaks Italian however we would go with the Southern Italian province.
We also developed a lot of the mapping technology from scratch to allow us to do some of the things you see there. We are very proud of the achievement and Alex, the wizard behind this task, is working on a geeky post to explain how we did it.
That’s all folks. We hope you enjoy the new maps. Happy clicking!
Pingback: Advanced Geolocation Filtering for social media mentions — THE UBER-BLOG