Displaying The Buffalo News’ #ecfare coverage
To cover the Erie County’s Fair food, The Buffalo News deployed a team of eight reporters to tweet pictures and locations of their food discoveries. I used two different systems, a categorized display featuring card elements as well as a map, to share the 75 tweets.
Although both displays pull from Twitter, they use two entirely different methods to do so. The map, amended from The Seattle Times’ mayday map, uses grunt as a task manager and node.js for the server. Although I had experienced both programs before, this project helped me familiarize myself as I tracked how the tweets went from the Twitter client to a SQL database before being served as JSON for the website itself.
On the other hand, the category display, adopted from Carla Astudillo’s Feats Thru Sheets, used IFTTT to pull in tweets from the journalists to a Google spreadsheet where another reporter would categorize each tweet as well as provide curation. Once the day arrived, I discovered my program didn’t catch the replies and the retweets so those were caught instead during the review process. Additionally, a 15-minute lag for the IFTTT recipes as well as a stressful unexplained hour delay provided frustrating.
In addition to IFTTT incidents, another problem was deleting a tweet from the map. Although the map was fantastic about not missing tweets, it also overridden any simple delete of a tweet from the JSON or SQL database. Eventually, I wrote another condition using the tweet’s id to a function that prevents retweets and replies from being added to the database.
Since the map is only able to locate a tweet given a precise location, the map is reliant on journalists turning that location on for each tweet. Additionally, the geolocation provided in the picture was off by a few blocks in certain instances.
If I did this project again for #ecfare 2017, I would adapt both programs to connect to Twitter’s API in the same way. Instead of double-checking after the IFTTT tweet delay, more focus could be on other elements of the project. I would also make it easier to catch tweets before displaying them on a map and also be able to edit the tweet coordinates as needed.
Other feedback included choosing a stronger contrasting map layer as well as making the markers easier to select on the map.
View the app’s and map’s code on GitHub.











