Announcement

Collapse
No announcement yet.

Text analysis or Qualitative Data Analysis

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Text analysis or Qualitative Data Analysis

    I'm looking for a program to help me dissect a collection of texts and mine it for data.

    I want the ability to
    -keep the original text
    -highlight different phrases or passages while reading, -assign them to categories / keywords.

    Prefferably the categories should form a hierarchy.

    Then I would like to be able to search, sort, crossreference the marked text according to my categories.


    Examples of use

    1) I actually want to use it to plan my trip to italy.

    I'm reading lots of stuff online and I would love to tag it with "city-name" "place name" and category (food/attraction) and then be able to browse something via tree.
    For instance, open "napoli"->"attractions"->"blah blah palace" and then I could read all the different comments i copied from all different sources, in one place.

    Or if I want to plan the trip, then I can get a report of only "traveling directions" from each site.

    2) suppose I want to follow the presidential campaign.
    I could save all different speeches, and tag excerpts with the speaker's name, and some tags according to an ontology of topics.
    Then if I want to retrieve all of McCain's quotes about the war in Iraq, I could.





    There are several web applications like Diigo or Evernote that do this, but I find them not as robust, and I want a desktop application.


    Apparently there's a whole pseudo-scientific methodology behind it that is called QDA (Qualitative Data Analysis).

    As expected, because it's a methodology that only exists in universities, all the desktop QDA applications are crap.They are all ugly, only work with odd free formats and are completely and utterly unintuitive.

    Can you help?

    Examples of the less shitty programs:
    Weft QDA - which is seriously under featured.
    Provalist Research QDA Miner - which is overly wierd, and filled with odd concept features.
    Last edited by Sirotnikov; September 6, 2008, 21:23.

  • #2
    I actually want to use it to plan my trip to italy.
    you really are thinking about the this too much if you want special indexing software to plan a trip
    Co-Founder, Apolyton Civilization Site
    Co-Owner/Webmaster, Top40-Charts.com | CTO, Apogee Information Systems
    giannopoulos.info: my non-mobile non-photo news & articles blog

    Comment


    • #3
      btw that Evernote yuo mentioned seems to have also an application for win/mac as well as for mobile.
      Co-Founder, Apolyton Civilization Site
      Co-Owner/Webmaster, Top40-Charts.com | CTO, Apogee Information Systems
      giannopoulos.info: my non-mobile non-photo news & articles blog

      Comment


      • #4
        yes i know
        i am not fond of it.

        Comment


        • #5
          no one has any ideas?

          Comment


          • #6
            As you say, there is an entire field around this ... we do somethings like this at my work, but i've never seen a general-public application that does what you ask. Evernote/etc. is probably the closest I could think of.
            <Reverend> IRC is just multiplayer notepad.
            I like your SNOOPY POSTER! - While you Wait quote.

            Comment


            • #7
              in that case, do you have a non-general-public application advice?

              Comment


              • #8
                Well, not in the sense of picking an application ... our stuff is all done custom, in ways that wouldn't help you at all.

                Oddly enough, my GF's dad was interested in something very similar, but he never found anything that would work... I considered writing it myself, but the I/O burden is just a bit too high for something that would not be written terribly efficiently (with my knowledge of the requisite programming aspects being average to below average), especially for PDFs which I don't know how to parse anyway.

                You could use any database software (Access, Filemaker, etc.) to accomplish something rudimentary ... just store each article in the database (as a 'blob' or giant text category, and/or as a link) and then type in the keywords (either as separate columns, or as just one generic keyword field). Either do it vertically (one row per keyword, possibly in a separate data table), or put in several keyword fields. Then write a simple form that looks up the keywords based on what you type in.
                <Reverend> IRC is just multiplayer notepad.
                I like your SNOOPY POSTER! - While you Wait quote.

                Comment


                • #9
                  I just read your topic, qualitative analysis and text, have you tried nvivo?
                  In da butt.
                  "Do not worry if others do not understand you. Instead worry if you do not understand others." - Confucius
                  THE UNDEFEATED SUPERCITIZEN w:4 t:2 l:1 (DON'T ASK!)
                  "God is dead" - Nietzsche. "Nietzsche is dead" - God.

                  Comment


                  • #10
                    just store each article in the database (as a 'blob' or giant text category, and/or as a link) and then type in the keywords (either as separate columns, or as just one generic keyword field).
                    doesn't help
                    i want to label parts of text - having a starting position and an ending position.

                    I just read your topic, qualitative analysis and text, have you tried nvivo?
                    no
                    it's a bit over the top.
                    it suggests building all sorts of different models to represent different ideas / concepts / moods, or what ever.

                    i merely want to build a simple topic ontology, and then process a growing colelction of texts, "highlighting" parts of the text and labeling them with different key words.

                    then i would like to be able to get reports of the highlighted portions according to the keywords / timeline / speaker.

                    Comment


                    • #11
                      Well then, have one database that has multiple data tables.

                      One data table is
                      (article number) (article as blob) (any article-level keywords)

                      and another data table
                      (article number) (startpos) (endpos) (keywords)
                      <Reverend> IRC is just multiplayer notepad.
                      I like your SNOOPY POSTER! - While you Wait quote.

                      Comment


                      • #12
                        Well, I tend to go with snoopy, if you look for those few functions, just do it yourself.

                        Or submit to the awesome power of nvivo
                        In da butt.
                        "Do not worry if others do not understand you. Instead worry if you do not understand others." - Confucius
                        THE UNDEFEATED SUPERCITIZEN w:4 t:2 l:1 (DON'T ASK!)
                        "God is dead" - Nietzsche. "Nietzsche is dead" - God.

                        Comment


                        • #13
                          do you use nvivo yourself?

                          Comment


                          • #14
                            Actually no, but I was supposed to, had I got myself involved with another project a year ago in another Uni. We were doing semantic data mining, text analysis and so forth. But that wasn't what nvivo was supposed to do, that was just one of the tools they used, we were actually developing such system that actually does this well, for hospitals (so a bit different). They used nvivo for all kidns of stuff, even analyzing interviews, because there were tons of them.
                            In da butt.
                            "Do not worry if others do not understand you. Instead worry if you do not understand others." - Confucius
                            THE UNDEFEATED SUPERCITIZEN w:4 t:2 l:1 (DON'T ASK!)
                            "God is dead" - Nietzsche. "Nietzsche is dead" - God.

                            Comment


                            • #15
                              semantic data mining, text analysis and so forth
                              what did you use?

                              Comment

                              Working...
                              X