Final Post – Reflection on my Digital Heritage Project ?

This is my last post on the class blog before the end of the unit. I’d just like to thank everyone for a wonderful semester, through the ups and the downs. There were things I got in an instant (success!) but there are also things that still baffle me, with more questions to be answered and possibilities to be explored!

But after everything I’ve learnt a lot and I’m excited to have had new experiences and to have gained new skills and knowledge.

Tim, thank for all your help and support with making the vision for my project a reality.

This is my longest post yet, so see the reflection on my project below the cut!

Over and out,


My aim for my Exploring Digital Heritage project was to identify new and exciting ways for digitised collections to become accessible to people through platforms already available. Early on I also aimed to incorporate my skills and knowledge from my journalism minor – something that I’m really passionate about, which so far has consisted of two units and a short internship with my uncle at the Herald Sun down in Melbourne – into the field of digital heritage. Luckily I was able to identify early on an idea for a project I felt passionate about, combining skills, knowledge, and ideas of my major with my minor to create something fun and engaging.

The plan was to create a Twitter bot inspired by the content of week two’s workshop. After looking at TroveNewsBot, TroveBot, and NYPLEmoji , I wanted to create something that combines the historical newsworthiness of TroveNewsBot with the fun and unique interaction of NYPLEmoji. Thus TroveEmojiBot was born!

My vision was to create a Twitter bot that responds with a news article from Trove’s digital collection when tweeted at with an emoji. Ideally, the corresponding article would relate to the emoji, either directly or in a broader conceptual sense. Targeted at young people using the social media site Twitter, the aim was to use to appeal of emoji create a fun and engaging way of encouraging young people from the ages of 18 to 25 to explore Trove’s vast digital collection though a familiar platform.

The bulk of the time working on this bot was spent editing the image.json file, the file that the bot refers to when sending out an article that corresponds to an emoji. Though I’m not sure how many emoji I used in total, I went about finding Trove articles for as many as I could. A rough estimate would be about 250 emoji. Initially I planned on using my journalism skills to assess articles on Trove for newsworthiness. The following are criteria for a newsworthy story:

  • Impact/Significance – a story that affects a great number of people
  • Proximity – the closeness to a story to the audience, can be geographical or cultural
  • Conflict – wars or dispute, but can also include sports matches, politics, and general rows
  • Human Interest – what people are doing, usually quite emotive stories
  • Novelty – anything unusual or doesn’t happen often
  • Prominence – stories involving well known people

I used this criteria in hand selecting the Trove articles for a number of emoji, however due to time limitations it was not efficient enough to meet deadlines. I then changed my strategy to search for the direct meaning of the emoji itself or a broader concept. Often the most relevant search had text errors from transcribing the content of the article, so I would always have to check before adding it to the json file. To keep the stories interesting I would add buzz words like ‘industrty’., ‘causes’, ‘havoc’, ‘championship’, ‘court’, etc. which allowed me to still find some stories of significance and newsworthiness to keep the bot interesting. Another issue I encountered in searching for articles to use was how short a number of articles seemed to be. To make the results more interesting I changed the search criteria to only include articles over 100 words. In addition, I would randomise the articles by occasionally limiting the search results to particular states, and would sometimes use this technique if I thought one state had a better chance of a more interesting story than another. For example, when searching for anything associated with deserts I would limit the criteria to results only from Western Australia or the Northern Territory.

This part of the project was the most interesting. I learnt things about Australian history I would not have known otherwise. In searching for results I learnt that horoscopes really took off in Australia during the 1950s, and have noticed changes in journalistic writing. Some headlines and stories I found were absolutely ridiculous, and I enjoyed sharing these findings with my friends and family. I hope that people using my bot have the same experience.


‘Computer Nerd Lockout’ –  One of my favourite headlines to date. Find this story here


‘Judge Rebukes Smirking in Court’ –  Filling in a niche for court reporting jokes on my personal blog. Find this story here!

Fun aside, not all emoji has a corresponding news article. Some emoji was just irrelevant, but a majority of the emoji that doesn’t have an article was due to time constraints and the difficulty of finding a related article. For example, the many variations of the family emoji were just too difficult to find due to the amount of elements within the emoji, for example a two mothers and a daughter.

When editing the json file my computer would use the wrong quotation marks, curly instead of straight. This took a while to fix as I was using text editor, but I was later shown a program by my partner called Sublime Text that allowed me to find and replace.

After editing this file, the next step for me was to test it by getting it running on my computer. In my project proposal I outlined that the biggest difficulty I would have with my project was getting it running and hosting it somewhere. I outlined two solutions to this problem; the first was to consult tutorials, while the second was to consult people with knowledge in the area. Tutorials gave me some guidance and a vague idea of how to carry out this second part of the project, however it was difficult to follow tutorials when my limited knowledge of coding meant that I didn’t know what a lot of the files meant. Because of this I sought the advice of Tim who guided me in the right direction to set up the API keys so we could get it running on a computer and eventually host it.

I wish I had the knowledge to get it running on my computer, but after one session in the lab with Tim the bot was up and running and I came away with a more solid idea of how the processes of getting it to run work.

Despite my lack of technical skill, I would consider the bot a success. Within an hour of the bot going live I already had people on Twitter, though not my intended audience, interacting with it, with thanks to Tim who spread the word through his own account. Most people currently interacting with TroveEmojiBot are from the heritage and library services sector, and I’m glad to see that they are enjoying it. Though it hasn’t quite reached my target audience – perhaps it will in time, I’ve already spread the word on my Tumblr blog – the sentiment of people exploring Trove’s collection and discovering something new, fun, and exciting still stands.

Because I chose an area I feel passionate about in heritage and journalism I plan to keep TroveEmojiBot running as an ongoing project. In order to keep the bot going a few things will need to happen. Currently the bot is running on Tim’s Heroku account. I’ll need to transfer the bot from his account to my own to be able to access and update it in the future. After watching Tim set the bot up, I have a general idea of where to go from here and feel somewhat confident in setting it up and transferring it over, though I will still need to seek some assistance either through tutorials, Tim, my partner, or a combination of the three. To keep the bot going from there will require new additions to the image.json file which I can do whenever I come across something interesting through my other units, or if I discover something new and exciting that I’d like to share looking through Trove independently of university.

All of these actions will not only help by keeping the bot going, but will also improve on my digital literacy skills. The end result is something that’s both exciting and practical which is coming away with new knowledge and a new skills set that I can use in the future for university, the workplace, or for my own personal projects.


Feel free to check my bot out @TroveEmojiBot

Leave a Reply