A beautiful project by Katherine, Gillian, and Erik.
We are creating a Twitterbot to engage with people interested in the Canadian publishing industry. Our bot will publicize the work done in MPub—as a stand-in for The Book of MPub, we will use our colleagues’ essays on the TKBR server to engage our audience. We have broken our process into three steps: “The Manual Work,” “The Ideas,” and “Coding and Testing.” A tentative fourth step (time permitting) will involve an audience analysis of our bot’s followers or related networks.
Step One: The Manual Work
Further to this, we brainstormed key topics and hashtags for our Twitterbot, and created a list of important publishers for our bot to be following, based on people we (as individuals, and the existing MPub twitter account) are already following and thus aware of.
We emailed the mpub-15 list for permission to share the essays published on TKBR, identifying key themes of each and tracking responses in a spreadsheet. This will help us to select particular #topics to focus on.
Step Two: The Ideas
Next, we brainstormed bot behaviours. These will include:
- following relevant accounts on the basis of their description (key terms in bio), who their followers are (ie. important publishers), or their behaviour (tweets about a topic)
- following back people who follow us
- retweeting particular topics and certain specific #hashtags
- tweeting @ new followers
- tweeting out links to new essays posted on TKBR; tweet new podcasts, newsletter
- tweeting old essays to people who:
- follow us and then mention the relevant #topic
- mention a relevant #topic enough times in a certain period of time (tbd)
- tweet @ us and mention the relevant topic anywhere in tweet
Step Three: Coding & Testing
In this step, we will implement, test, and assess the utility of each bot behaviour. Independently, individual group members will use their own Twitter account to write and test each piece of code. Once we’re confident that it works, we will compile it into a single master document. This document will also include code necessary to keep our bot working within Twitter’s rate/access limits and run smoothly over a longer period of time. Ultimately, we will upload the bot to the TKBR server.
Our deliverables will include:
- our rough work (such as completed spreadsheets)
- a specification sheet detailing bot behaviours and an assessment of their functionality (such as effects on audience growth and engagement)
- our report, including:
- our decision-making process for deciding topics and behaviours
- documentation of our development and learning process
- time permitting, an audience analysis of our Twitterbot’s followers