One Million Tweet Map

Link to visualization:


The One Million Tweet Map is a social media tool which helps visualize the past one million tweets obtained from Twitter. It updates real time with the older tweets getting deleted as newer ones are updated. The product was built by Maptimize and uses OpenStreetMap, Leaflet and CarloDB.

This visualization helps in understanding where various tweets pertinent to a particular topic are coming from. It features various tools as discussed below, the most important of which is the hashtag based search. There was not extensive documentation on the visualization, but the developers say that they use the Twitter Streaming API to collect the tweets. The API allows for extracting the past one million tweets from Twitter, which is the reason why they restricted the visualization to one million tweets as well.

This usage of this website can range from a casual audience who just want to play around with the visualization. It can also be useful for a more serious audience who would want to analyze sentiments of people related to a particular topic/hashtag.

The sidebar panel where the various inputs can be chosen by user is displayed below:

The interface is friendly and not confusing. Hence, user should not face any problems while changing the various metrics.


  1. Heatmap Layer
    It shows a heatmap on the world map indicating where most tweets are coming from. The developers could have displayed a scale correlating the intensity of the color and number of tweets, but apart from this, the heatmap is decent enough.

  2. Cluster layer

    This view shows a lot of clusters (shown as circles) indicating the number of tweets coming from that part of the map. This can be zoomed in which causes the clusters to break into smaller clusters and finally it shows the individual tweet when no other tweets exist in that region on the map. The below image shows the default zoom for the map and shows various clusters around the world.
    As you zoom in more, you see that the clusters keep breaking as shown below:
    There is a sentiment View in the rendering which shows the different clusters in three colors: green(positive), yellow (neutral), red(negative)
    Near the zoom control, the number of points currently on the screen is also shown. This is very useful for finding the percentage of positive, neutral and negative tweets. Note that this analyses sentiments only for tweets in english.

    The colors used are intuitive and helpful for understanding how the sentiments are for a particular hashtag (using hashtag search explained below).

  3. Day/Night layer

    This view shows the regions of the world which are currently in the day and the regions shaded in black are currently in the night.

  4. Time filter Three time filters are given in the website:

    Therefore, we can see the user is presented with all options in a friendly interface while at the same time giving the flexibility for viewing tweets for a variety of time periods.

  5. Top 5 Countries
    This shows the top 5 countries from where tweets are coming from. This is useful when you are searching for a particular hashtag and you need to figure where most of the tweets are coming from.

  6. Hash-tag based search
    The search input field at the top allows the user to search tweets based on hashtags. This is one of the best features of this visualization where the tweets being displayed are restricted to a particular hashtag. Example search is shown below:

    What could be improved:
    It was a bit confusing as to what the red flashing dots on the screen are. This is because the red color dot is also used to show negative sentiment. So, I was not sure initially if the red flashing dots were new tweets or they were negative tweets. However, after reading the documentation, I found that it just meant new tweets. The visualization could have used a better legend to indicate the heatmap also.