We are a small business with a 5-year GSA Schedule good through 2024. You can email the founder and CEO Dr. Stuart Shulman (stu@texifter.com) or find us on Twitter @DiscoverText.
DiscoverText Workshop Feedback
“Being able to sort by meta-data, being able to embed tweets (esp with video, pictures, emoji) in a platform where they can be coded is incredibly useful. I’m also excited by the ability to see replies and who someone is responding to. I’m very focused on civic dialogue and argumentation, so on additional feature would be to be able to collect in a threaded format (to see whether people are engaging in a sustained vs. one-off manner). Also, finding a way to collect Reddit data would be amazing. Having a tool created by social science researchers is so incredibly useful.”
“Easy and navigable software.”
“I have previously worked on different open source softwares to conduct Twitter sentiment analysis for research purposes. However, after attending today’s session, I really found the discovertext software much more extensive and to be offering so much more than other platforms. I really liked the reply_to meta data feature, and the facility to collect multilingual data really offers a greater bandwidth for research purposes.”
“The user interface seemed very easy to navigate which was a plus. I have mainly focused on dealing with the twitter api directly, so this software seems like a great way to interact directly with the data.”
“I like it because of multilingual interface!”
“Very informative workshop! Generates so many ideas for future research! I really appreciate that this software maintains Twitter compliant datasets which makes it safe to use in academic research. As I mentioned during the workshop, it would be interesting to see statistics on # of Tweets deleted/suspended from previously collected sets.”
“It’s a great software with a friendly user interface. I find it interesting, especially because it requires little knowledge of computational skills.”
“The workshop was brief and informative enough for starters. I was happy to see that there is the opportunity to work with different languages.”
“Very helpful to see some of the software’s functionality demonstrated, such as filtering data by influencer score. I would love to see some additional resources that describe or demonstrate these types of options. It was also very helpful to see the quick demonstration of machine learning at the end. This is also something I would like to know more about.”
“The most useful feature for me is the bucket and dataset feature. I can filter and cluster the data, explore the trends of the current discussion while also search for different voices that can appear in the Twitter discussion. It will be interesting to learn more about further potential to explore the bucket and to classify the dataset in a more advanced-level workshop.”
“I really liked the practical demonstration that was grounded in empirical cases. It was also very rich the fact that Dr. Shulman shared good and bad experiences with the use of the tool. Another important point was the discussion on ethics. I’ve used other tools, such as Gephi and Nodexl tweets import plugin, but the loss of context of tweets is really a big problem of these tools. I believe that discovertext meets this need.”
“It seemed easy to use for new researchers and led to new insights about twitter data easily. It looked intuitive.”
“I like the ease with which we can collect Twitter data.”
“This is an easy way to compile the kind of Twitter data I’m looking for for my research. Other tools I’ve used require a much higher skill level and/or are not as easy to search through.”
“Faster coding.”
“With DiscoverText, one can use advanced search filters (eg. specified metadata fields) to classify searches and removes duplicates. This could be a powerful tool for researchers when gathering survey data that has open-ended questions, help saving time and effort. Further, DiscoverText can generate reports in RTP, XML, CSV format that could enable inter-coder reliability to increase accurate analysis. Lastly, being able to cluster information, helps to work on a large amount of data more manageable as DiscoverText can show the main trends in them quickly. With DiscoverText, there are restrictions from Twitter where it is limited to the number of tweets being exported per day. On the other hand, this feature is supported by the recent integration with NodeXL, which is also an interactive network visualization and analysis tool, providing the ability to export directly to a format that can be supported by NodeXL.”
“Ability to code a large data set quickly seems to be the key strength.”
“It has a variety of functions and filters that would be useful for research projects. I think having a lot of filters and different ways of extracting subsets of original data is very unique (A software for Twitter collection that I used a little bit does not have such features.). The workshop covers many important things.”
“Sparked a new flow of ideas for data collection and the possibilities for future research. This is my first time seeing data collection on social, so I cannot personally compare. I really enjoyed the workshop, I personally would like it to be longer and include multiple examples of topics and associated buckets. Exploring more of the do’s and dont’s, to avoid user error.”
“The ease of use – I had taken a day long workshop of using R for this and was overwhelmed. This appears to be much more user friendly.”
“Looking forward to using the software to conduct qualitative research.”
“I love the data collection, filtering, and meta data exploration features. I’m sure that, given a little time to explore it, I will find about 100 more things to love!”
“I found the software really interesting. I have previously collected and analysed tweets with DMITCAT, Ncapture & Nvivo & Maxqda. I like the fact that Discovertext gives access to visuals and not just the verbal part of the tweets – as that is a major component of Twitter-communication. I think the analytical possibilities in the software seemed cool, and the software somewhat easy to learn.”
“This workshop is very useful for me. Now that I knew there exist software that is going to make my research easier.”
“I have used other tools to collect twitter data but they do not have the same abilities to code or classify that DiscoverText has. I have also found that to view a tweet from my other dataset tools that I have to leave the tool and view each tweet natively within Twitter which makes it much slower for viewing the individual tweets. When I previously used DiscoverText it was self-taught and took me a long time to work out everything that you demonstrated in less than an hour – I think this introduction would have saved me a lot of time experimenting on my own.”
“This is my first software package for twitter research. From what I saw, being able to look at pictures and videos while gathering data will be extremely helpful.”
“I have to say that DiscoverText seemed like an easy to use but hard to master tool, which is great :)”
“I find discovertext very interesting. I have used atlasti v8 with twitter but the findings are not so accurate and it lacks of complete options for research. I would like to try the software because it looks so much robust for very specific search. Dr. Shulman’s explanation was nice and complete, thanks for answering my questions!”
“This software is an extremely powerful and versatile tool for examining one of the most important data sources around at the moment. It is useful for qualitative and quantitative approaches, and can result in often unexpected insights into the nature of people’s online identities and networked interactions.”
“This was a brand new introduction to DiscoverText for me, which I’ve never used. I have used other Twitter-gathering tools like NodeXL and TAGS, so I had some understanding of how tweets are pulled and how to analyze them. It was great to see all the tools available within this software that normally I would have to use multiple tools for (e.g., coding, visualization). I still need to play with it a lot more, particularly as this is not my primary way of researching social media, so I get a little overwhelmed by what to do with all the information available to me. Thank you!”
“This was fantastic. I have used DiscoverText before several years back (2015-2016), and was pleased when I got the email about the workshop. This was exactly what I needed to refresh my memory about what DiscoverText can do and how to use it, and I am now reinvigorated to use DiscoverText again in my research!”
“I like the easy way to explore unstructured data. This is something that can be very tedious. Also the possibility to build a dataset for exploring in a network tool like gephi. I think that is a way to find insights that could be difficult to obtain other way.”
