Friday, April 1, 2022

CasualConc goes 3.0!

After 3 and half years of no activity on this blog, I am happy to announce a new release of CasualConc (though it is still a work-in-progress).

This version has some new features and bug fixes as well as interface changes.

Two major additions are basic XML file handling and grammar search using dependency grammar tags with Stanford CoreNLP.

The XML handling feature is fairly basic, but CasualConc analyzes XML structure and you can select the tags or attributes to use for grouping/filtering.


You can use filters in tools.


Grammar search utilizes dependency grammar tags assigned by Stanford CoreNLP. You need to install Java and Stanford CoreNLP separately (the instruction is on a new manual).

You can use this new feature on any of the languages Stanford CoreNLP can handle by specifying dependency grammar tags.


For English, you can select a pre-specified grammar structure for search.

I prepared a sample grammar database file (with Sherlock Holmes novels/shorts) on CasualConc site, so you can feel what this tool is like.

There are other more minor additions, such as filtering files for analysis and lexical diversity measures on Word Count/File Info.

This is still a beta version, so there may be many issues. Please report any problem you encounter. That will help improve CasualConc.

In any case, please enjoy playing with grammar search. It's fun!

Sunday, September 23, 2018

CasualConc 2.1 is released with an updated manual!

I just released an updated version of CasualConc along with an updated manual.

It has many new features and a lot of bug fixes.

Major changes are an addition of Frequency Plot tool. Now you can draw box plots, violin plots, bean plots, histograms and bar plots with mean frequencies of items (different from bar plots of Chart tool). Also you can plot relative position of items.

A very simple batch tagging function is also added. If you have TreeTagger, Stanford CoreNLP, or MeCab on your mac, you can create tagged texts on CasualConc.

There might be some bugs remaining, so please try it and report any bug, problem, or feature request.











Sunday, May 8, 2016

Current status of CasualConc

I haven't written here for quite a while, but I've been fixing bugs on CasualConc 2.0. Nothing fancy has been added since I posted an entry here on this blog. A few of the new or enhanced features are:

Visualization tools:
label coloring is easier; you can grab label colors from Label Coloring panel or even directly from the Label Coloring Database.

File Info:
tag filtering is available in File modes (simple/advanced); you can set multiple tags to count frequencies of words in different tagged parts; for the first time, you can set the tagged area using XPath (experimental)

This is an example of TF-IDF using XPath option. Speakers are tagged on the Friends scripts and simply frequency counts by each character's utterances are used to create this. If you know Friends, you might be able to see something meaningful.



I've been playing with XML to implement some basic XML features, but I'm not sure how to create a good interface. Any idea?


Finally, this is a kind of important note. If you have used R on CasualConc and recently upgraded R to 3.3, you need to turn off/on the Use R for Visualization on Preferences. It looks like you need to re-install non-default packages after you upgraded to 3.3. I got one report that even after re-installing packages, some packages were not installed properly (not the packages CasualConc tries to install, but the packages those 'packages' depend). In that case, you need to manually reinstall those packages on R or RStudio. I'll see if I can add a function to check them or at least display error message from R (I haven't figured this out yet...).

Sunday, October 18, 2015

New feature showcase

This is just a series of images that showcase the new features of CasualConc 2.0.

Line Chart

Line Chart
Bar Chart

Pie Chart

Radar Chart

Scatter Plot

Word Cloud (frequency)

Word Cloud (TF-IDF)

Hierarchical Cluster Analysis (dendrogram)


Correspondence Analysis

Principal Component Analysis (factor scores)

Principal Component Analysis (factor loadings)

Multi-dimensional Scaling (MDS)

MDS (3D)

Network Analysis

Network Analysis

Concordance Plot

Random Forest (keyword extraction)

Regular Expression Practice

Vocabulary Profiler

Tokenization (Japanese)

Tokenization (Korean)

Tokenization (Chinese)

Tokenization (Thai)




Sunday, October 4, 2015

CasualConc 2.0 beta

I'm pleased to announce the immediate release of CasualConc 2.0 beta, a new version of CasualConc.

For the last couple of years, I've mostly worked on CasualTranscriber, which is now version 2.5. Now it has most of the functions I wanted to use. So I decided to update CasualConc.

CasualConc was first released in 2008 and has been constantly updated up until a couple of years ago. The release of OS X 10.9 Mavericks changed something in the developing environment and made me unable to update the project. This was partly due to the scripting language/bridge I used to write CasualConc.

This new version of CasualConc is almost entirely rewritten. It is written in RubyMotion, which is kind of a successor to RubyCocoa.

Due to the difference in regular expression engine, the speed of text processing is a bit slower with RubyMotion, it is compensated by parallel processing. Also due to the same reason, your search result might be slightly different from the one with the older version. This is because the two engines treat some symbols/marks differently.

This new version may not have all the functions you use with 1.9.x, so if you need a particular feature, please let me know. I'll see if I can bring it back to the new version.

CasualConc 2.0 has the following new and improved features:

  • text mode
  • opening a result window for Collocation and Word Count
  • much faster File Info tools
  • collocation frequency table
  • utilization of build-in tagger and integration of TreeTagger (including an installer)
  • integration of a statistical environment R (needs to be installed separately)
  • keyword extraction with Mann-Whitney-U and Random Forest (requires R)
  • visualization of frequency information (requires R)
  • and more...

A manual, which covers most of the features, is available and help is included in the application.

CasualConc 2.0 beta is compatible with Mac with OS X 10.9 Mavericks or later, including 10.11 El Capitan.

It is very likely that this version still has many problems. Please try it and if you find any problems, your feedback is greatly appreciated.

This slideshow (in Japanese) presents new features. Text is minimal (mostly screenshots).