You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 24 Next »

Short Tutorial

Voyant

nGrams

Bookworm

 

Group 1 - Voyant

Tips

As a group explore and discuss

  • Spend some time exploring the interface.  What can you tell about the underlying data as you explore with the tool? 
    • Is the text clean? Indexed? Filtered? Anything else interesting you note about the data?
  • Explore the functions of the tool.  Attempt to make claims about the intellectual content of the text based on the tool and its visualizations. (Feel free to reach a little.)
    • Try constraining the various tools to specific terms or phrases.  Do any words trend together? Inversely?
    • Can you isolate phrases of interest?
    • Identify and operate the sliders - in what ways could these be useful?
    • Adjust the stopwords to dampen "noise" and heighten "signal"; what are potentially positive and negative consequences? 
    • Do any of the the visualizations help illustrate what you notice/assert regarding the text? Are some misleading? Confusing?
  • Consider the value of the tool
    • What can you manage to do?  What is this tool good for?
    • What sorts of things did you want to do, but could not?
    • What can you infer from the interface about the text? What is still opaque?

After your exploration, be prepared to report your findings to the other team(s) and take their questions.

Group 2 - nGram Viewer

Tips

  • Any browser should work fine.
  • Your exploration can begin here, but you can also develop other interesting explorations. NOTE: you may have to hit the blue "Search Lots of Books" button to bring in the visualization. 
  • Help files for customizing the tool are at https://books.google.com/ngrams/info
  • Discussion of underlying data is here.  
  • Choose roles of recorder and presenter

As a group explore and discuss

  • Spend some time exploring the notes about the underlying data.  What kind of text data is this? 
    • Is the text clean? Indexed? Filtered? Structured? Anything else interesting you note about the data?
  • Explore the functions of the tool.  Attempt to make claims about the intellectual content of the text based on the tool and its visualizations. (Feel free to reach a little; definitely refine and make the input better.)
    • How comprehensive are these terms?  Can you make them more comprehensive by making them case insensitive?  What happens to your results?
    • Can you restrict the case of some phrases and make others case insensitive? How?
    • Can you segment by language to indicate data from Great Britain as distinct from US English?  What does it mean to find these terms in Spanish?
    • Can you add related terms?  Do these terms show interesting trending, either coincident or inverse? 
    • Add the coalesced term "(male+(chauvinism+chauvinist))".  Note the way the frequency of this form obliterates the other waveforms.  What can you do to temper this action? 
    • What sorts of supplemental data would be helpful in making sense of these visualizations? 
    • Play around - try other searches and customizations, observe and evaluate their effects.
  • Consider the value of the tool
    • What can you manage to do?  What is this tool good for?
    • What sorts of things did you want to do, but could not?
    • What can you infer from the interface about the text? What is still opaque?

After your exploration, be prepared to report your findings to the other team(s) and take their questions.

Group 3 - Bookworm

Tips

As a group explore and discuss

  • Spend some time exploring the notes about the underlying data.  What kind of text data is this?
    • Is the text clean? Indexed? Filtered? Structured? Anything else interesting you note about the data?
  • Explore the functions of the tool.  Attempt to make claims about the intellectual content of the text based on the tool and its visualizations. (Feel free to reach a little; definitely refine and make the input better.)
    • Add other communicable diseases of interest.
    • Try comparing Malaria as found in publications of the United States or the United Kingdom. 
      • Where does the data to facet in this way come from?
      • Can you make a narrative out of this graph?
      • What sorts of supplemental data would be helpful in constructing this narrative?
    • Conjecture as to what Class, Subclass and Narrow class mean.  Where would this faceting data come from?
    • Click on a spot on one of the plotted curves (and wait, rendering can take a little time).  What is this data in the drop down?  Explore it - how might it be useful?
    • Operate the date sliders.  What happens to the representation of the data when you zero in on certain years?  How does that affect the narrative you would tell about the trend of the frequency of a word?
    • What sorts of supplemental data would be helpful in making sense of these visualizations?

    • Play around - try other searches and customizations, observe and evaluate their effects.
  • Consider the value of the tool
    • What can you manage to do?  What is this tool good for?
    • What sorts of things did you want to do, but could not?
    • What can you infer from the interface about the text? What is still opaque?

After your exploration, be prepared to report your findings to the other team(s) and take their questions.

Bringing it back together

  • What sort of conclusions/suspicions can we draw from canned tools in general?
    • In what types or phases of projects would these tools be useful?
    • In what types or phases of projects would we need more control over analysis?

 

 

  • No labels