DiscoverText

Collaborative text analytics for academics, companies & govt

About DiscoverText

DiscoverText is a cloud-based text analysis and data science software used to evaluate large amounts of unstructured free text, survey responses, Twitter data, public comment to government agencies, and more. The semi-structured workflow allows users to create adaptive, custom text classifiers using machine-learning and crowd source coding to find relevant items, and sort them into categories such as topic or sentiment.

DiscoverText includes a number of multilingual, data science, text mining, human coding, and machine-learning features which help to evaluate large volumes of unstructured free text, and conduct precise analysis at scale. Users can create customized, reusable machine‐learning text classifiers or 'sifters' using the uClassify web service in order to accurately collect, clean, and analyze relevant text data. Results can then be sorted into topic, language, gender, sentiment, mood, and various other categories. The solution's combination of data science methods and tools for eDiscovery text analytics help to shorten the process, along with features for crowdsourcing. A collaborative annotation system and adjudication methods help to improve machine-learning by ranking human annotators over time. DiscoverText also includes a patented CoderRank approach which focuses on algorithmic techniques comparable to 'PageRank' for Google search, but tailored for large-scale text analytics.

Users can take advantage of features such as Boolean defined search, interactive word clouds, n-grams, and custom mini topic dictionaries, as well as filter items by date, geography, influence, networks, place name, and a variety of other metadata values. Features for deduplication and automated clustering of similar text allow users to create groupings to further accelerate the training of machine classifiers. Custom, complex search queries and human coding schemes can be saved for reuse at a later date.

Key benefits of DiscoverText

  • DiscoverText's CoderRank tool focuses on algorithmic techniques comparable to "PageRank" for Google web search, but tailored for large-scale, collaborative text analytics through enhanced machine-learning.

  • Eliminate duplicates and cluster similar text and comments with a click using DiscoverText, and auto-highlight any unique or offensive language.

  • Generate and drill into word clouds to visually understand text data and build mini topic dictionaries. Generate reports in PDF, XML, RTF, or CSV format for further analysis.

  • Form peer groups and project networks to collaborate on assignments and tasks within DiscoverText, and share or archive completed projects online.

  • Search for key concepts and text with Boolean defined search, and archive social media content from both free and premium Gnip Twitter data streams.

  • Devices

    Business size

    S M L

    Markets

    United States, Canada, Australia, China, India and 4 others, Japan, Germany, Brazil, Mexico

    Supported Languages

    Arabic, Chinese (Simplified), Danish, Dutch, English and 6 others, Finnish, French, German, Hebrew, Italian, Spanish

    Images

    Create and save complex search queries, human coding schemes, or just filter on active-learning machine classification scores
    Deduplication and automated clustering of near-duplicates gives users a high level sense of the data landscape
    Manage datasets and view the number of units, codes, coders, and coding percentages
    The uClassify feature offers interactive machine classifier histograms which allow data science teams to identify the items in a collection that add the most value when coded by humans
    View 5 more
    DiscoverText screenshot: Create and save complex search queries, human coding schemes, or just filter on active-learning machine classification scores DiscoverText screenshot: Deduplication and automated clustering of near-duplicates gives users a high level sense of the data landscape DiscoverText screenshot: Manage datasets and view the number of units, codes, coders, and coding percentages DiscoverText screenshot: The uClassify feature offers interactive machine classifier histograms which allow data science teams to identify the items in a collection that add the most value when coded by humans

    Features

    Total features of DiscoverText: 21

    • API
    • Automated Scheduling
    • Collaboration Tools
    • Commenting
    • Content Filter
    • Customizable Templates
    • Data Extraction
    • Data Filtering
    • Data Import/Export
    • Document Management
    • Duplicate Detection
    • Filtered Views
    • Full Text Search
    • Keyword Tracking
    • Metadata Management
    • Permission Management
    • Projections
    • RSS Feeds
    • Reporting & Statistics
    • Search Functionality
    • Summary Reports

    Alternatives

    Gavagai

    4,3
    #1 Alternative to DiscoverText
    Gavagai is a text analytics tool designed to help businesses of all sizes collect and analyze unstructured data using...

    OpenText Magellan

    5
    #2 Alternative to DiscoverText
    OpenText Magellan is a cloud-based machine learning & predictive analytics platform, which helps businesses analyze...

    Thematic

    4,9
    #3 Alternative to DiscoverText
    Thematic is a feedback analysis platform which pulls multi-channel feedback into a single view, discovers emerging...

    Mozenda

    4,4
    #4 Alternative to DiscoverText
    Capture web data, automate processes and scale. Trusted by thousands of businesses and over 30% of the global Fortune...

    Reviews

    Overall rating

    4,6 /5
    (107)
    Value for Money
    4,6/5
    Features
    4,5/5
    Ease of Use
    4,2/5
    Customer Support
    4,5/5

    Already have DiscoverText?

    Software buyers need your help! Product reviews help the rest of us make great decisions.

    Write a Review!
    Showing 5 reviews of 107
    Wasim A.
    Overall rating
    • Industry: Higher Education
    • Used Weekly for 2+ years
    • Review Source

    Overall rating

    • Value for Money
    • Ease of Use
    • Customer Support
    • Likelihood to recommend 10.0 /10

    Superb cloud-based software tool with powerful text analytics with social media integration.

    Reviewed on 2017/12/19

    It allowed me to conduct research that otherwise would not have been possible including both...

    It allowed me to conduct research that otherwise would not have been possible including both industry and academic projects.

    Pros

    The ability for those from the social sciences to be able to import and/or retrieve social media data, including historical data from Twitter, and analyse the data in order to answer research questions. This is because other tools may require a computer science background. Consequently, DiscoverText has been used in answering important social science questions leading to peer reviewed outputs. Over the years I have seen many tools appear and slowly wither away, however, DiscoverText has stood the test of time and has been growing in popularity.

    DiscoverText is not limited to academic uses and has a number of neat uses in the commercial world. A useful feature applicable to the commercial domain is the ability to retrieve and/or import data from Twitter and identify influential Twitter users, with the additional ability to use machine learning to sift influential users into different groups. For example, a football club may be interested to find out whether influential users are fans of the club or whether it is opposition fans causing a storm. To the best of my knowledge, no other tool is capable of doing this with this level of accuracy.

    Cons

    To be fair this is not a limitation of DiscoverText per se, as this is a restriction from Twitter, but there is a limit to how many units of tweets can be exported per day. This is not a major issue because there are enough features in DiscoverText that you may not necessarily need to export the data. This is particularly true with a recent integration with NodeXL which provides the ability to export directly to a format supported by NodeXL.

    Christine A.
    Overall rating
    • Industry: Environmental Services
    • Company size: 1 001-5 000 Employees
    • Used Daily for Free Trial
    • Review Source

    Overall rating

    • Value for Money
    • Ease of Use
    • Customer Support
    • Likelihood to recommend 10.0 /10

    A much better way to scrape data than learning how to code an API

    Reviewed on 2018/12/29

    I'm a PhD candidate who straddles the Humanities and Social Sciences, so I use DiscoverText as a...

    I'm a PhD candidate who straddles the Humanities and Social Sciences, so I use DiscoverText as a research tool.

    Pros

    My review of DiscoverText is a bit limited because I'm really only using it for Twitter. That said, the features are incredible. I know a bit about coding, but the prospect of learning JSON to use Twitter's API was doable but daunting. When I came across DiscoverText I was so pleased to find a way to search, use, and categorize Twitter data that made sense and would save me A LOT of time. I didn't anticipate getting access to so much useful metadata that was easy to navigate and use, so I was pleasantly surprised. The built-in bucket and dataset features are great ways to organize the massive amount of Twitter data that can be collected. The ability to code the data with peers within DiscoverText is also super useful. I really can't exaggerate how many features DiscoverText has that I didn't think I would need but have used to improve the quality of my scholarship.

    Cons

    The software has many features that I didn't find on my own, so the UI could be improved a bit. That said, the one-on-one tutorial that the founder provides helps mitigate this issue. The tutorial videos are helpful too! You'll just have to be prepared to set aside a few hours to really learn the program.

    Response from Texifter

    Dear Christine,

    It is really hard to express how inspirational a review like this is. You have really made our day. We are looking at 2019 trying to decide if this is the year to build v2 of a 9 year old interface. Thanks for embracing buckets & datasets; this was a tough sell to some folks over the years, but they are critical to User success.

    We are very grateful you took the time to write this generous review. Please write us if we can do anything for you.

    Thanks,
    ~Stu

    Katarina A.
    Overall rating
    • Used Daily for 1-5 months
    • Review Source

    Overall rating

    • Value for Money
    • Ease of Use
    • Customer Support

    Great for analyzing social media data- just not offline documents.

    Reviewed on 2018/04/07

    Pros

    If you need to analyze data from social media and survey monkey- it's a great tool. You can search for content by keywords and the data drops in per the chosen frequency. It breaks down keywords and phrases to a list in order of use- where you can drill into each word or phrase to see where it's used and also toggle between different ways of displaying the results.

    Cons

    I got the trial version to see if it would suit my purpose; I required a tool to analyze and cluster data from articles and other sources but just couldn't get it to work. If the trial period had been longer than 3 days ( I thought I signed up for 30- it's not very clear) I might have had time to figure it out. Better instructions would have helped. The instructions tell you what the features are, not why you need to use them which is not helpful for novice users.

    Response from Texifter

    Katarina,

    Sorry for the confusion about the length of the free trial. It was 30-days for many years and we changed it to 3 only recently. Please send a request to [email protected] and I will send you a 6-month license.

    For details about the features, we suggest you review some of the support materials:

    https://texifter.zendesk.com/hc/en-us

    As to why use the tools, perhaps review the tutorials:

    https://discovertext.com/tutorials/

    You might also find some answers as to why use the tools here, in the 200+ academic citations of the tools:

    https://discovertext.com/publications/

    Finally, I am available for 1-1 web trainings:

    https://calendly.com/discovertext

    So, I think you may not have fully tried to use all the customer support options. We work very hard to make sure newcomers get comfortable quickly.

    Stu

    Hossein K.
    Overall rating
    • Used Daily for 1-5 months
    • Review Source

    Overall rating

    • Value for Money
    • Ease of Use
    • Customer Support
    • Likelihood to recommend 10.0 /10

    Honestly,I can say DiscoverText makes analyzing social data not only easier, but also more...

    Reviewed on 2018/01/01

    It provides me more opportunities for working on my projects. Using it, I have access to many ways...

    It provides me more opportunities for working on my projects. Using it, I have access to many ways for doing research on social media data which have not before.

    Pros

    First, it is so easy to learn and use. Moreover, the DiscoverText founders provided some helpful tutorials and educational videos which are so handy and helpful. This software allows users to makes several datasets of one project. This enables a researcher to work on multi-dimentions of a certain project needless to create different ones. Furthermore, you can create a sample of your data very easily by making a dataset. Its buckets are very interesting also. Additionally, you can make some clouds of data by using cloud explorer feature. Finally, Clustering option is great! it makes working on big data easy and shows the main trends in them quickly.

    Cons

    As I can say, sometimes users may get confused by many links and pages. So, maybe finding what you want becomes difficult and you have to try some ways. Another con, in my point of view, is the obscurity of metadata meanings and algorithms. I cannot understand what some of the means and how they are calculated. Furthermore, I think some of metadata can be presented in some more useful ways. But at all, I should confess the metadata explorer is a great ability!

    Response from Texifter

    Thanks Hossein for an excellent review. We are preparing a new blog post now with a Metadata Dictionary for Gnip Twitter data. We agree that some of the fields are a bit confusing and we hope this new blog post will make the meaning of some of the fields more transparent.

    Jim W.
    Overall rating
    • Industry: Higher Education
    • Used Weekly for 2+ years
    • Review Source

    Overall rating

    • Value for Money
    • Ease of Use
    • Customer Support
    • Likelihood to recommend 10.0 /10

    We have been using the software to study twitter conversations on immigration going back to...

    Reviewed on 2018/01/29

    Easy data capture

    Easy data capture

    Pros

    It's ability to capture tweets and now the capacity to export to NodeXL gives us two tools that we use together to study the content and structure of immigration conversations onTwitter.

    Cons

    At times navigating the menus is counter-intuitive as is some of the terminology. Archives, buckets, datasets all kind of run into one another.

    Response from Texifter

    Dear Jim,

    Thanks for your and the generosity of your Tweets. We really appreciate that!

    I'd like the opportunity to visit GMU to make the case that archives, buckets, and datasets are essential parts of the text analytics methods we have engineered. Please email [email protected] if you would like to host a free workshop.

    Briefly:

    - Archives are raw data.
    - Buckets are subsets of raw data.
    - Datasets are coded by humans.

    Most projects proceed from 1 or more archives, to many buckets, to a series of codeable datasets. For example:

    - Collect 100,000 #metoo tweets
    - Deduplicate the archive
    - Create a bucket of seeds and singles
    - Search the bucket for key terms
    - Create a new bucket with results
    - Create a dataset and code it for relevance
    - Train a relevance classifier
    - Apply the classifier to new archive samples
    - Repeat as needed

    The key point is that raw data is messy in the archives, cleaner in buckets, and fully refined and classified in datasets.

    Showing 5 reviews of 107 Read all reviews

    DiscoverText FAQs

    Below are some frequently asked questions for DiscoverText.

    DiscoverText offers the following pricing plans:

    • Starting from: US$99,00/month
    • Pricing model: Subscription
    • Free Trial: Available

    PROFESSIONAL: $99/month ($49 for academics, $24 for students) ENTERPRISE: $2,000/month ($1,000 for academics, $500 for students)

    DiscoverText offers the following features:

    • Advanced search
    • Attach memos to documents, datasets and archives
    • Boolean defined search
    • Boolean queries
    • Bucket filtered documents and search results
    • Buckets
    • CloudExplorer
    • Coding tools
    • Comparison tools
    • Connect and work with peers via web browser
    • Create and save complex search queries
    • Custom topic dictionaries
    • Data exports
    • Document filtering
    • Duplicate detection
    • Filter by metadata and threshold classification
    • Generate high-level summary and detailed reports
    • Highlight unique text
    • Human coding scheme
    • Keyword search
    • Language detection
    • Live data feeds
    • Machine-learning classifiers
    • Measure inter-rater reliability and validate results
    • N-grams
    • Offensive language detection
    • Re-use models
    • Reporting tools
    • Schedule repeat fetches
    • Taxonomy classification
    • Text analysis
    • Topic clustering
    • Word clouds
    • uClassify - Machine learning web service

    DiscoverText has the following typical customers:

    Small Business, Large Enterprises, Public Administrations, Non Profit, Freelancers, Mid Size Business

    DiscoverText supports the following languages:

    Arabic, Chinese (Simplified), Danish, Dutch, English, Finnish, French, German, Hebrew, Italian, Spanish

    We do not have any information about what devices DiscoverText supports

    DiscoverText integrates with the following applications:

    SurveyMonkey, Twitter

    DiscoverText offers the following support options:

    Phone Support, Knowledge Base, Video Tutorials, Online Support

    Related categories

    See all software categories found for DiscoverText.