- Slides: 25
Visualizations Follow-up. Data: dimensions. Basic graphs. Gapminder. Excel. Wordle. Homework: Postings: research and experiment.
Follow-up from guest lecture • Summarize? • ?
Natural Language Processing • Text to – text or – data extraction
Speech recognition • input is sound wave captured as electrical amplitudes and frequencies • first step is doing Fourier Analysis to encode sound wave as set of sine waves • next step is segmentation: divide into (best guess) at syllabus (phonemes) • THEN use statistical approach and/or grammar approach to make matches
NLP and Speech reco • Can combine • For each, issues of scope, speaker(aka user), training or not • Distinct applications – – – limited (directed) speech dictation data extraction translation ? ? ?
Consider • several possible topics for presentation • I will demonstrate duolingo. com which I am using to learn Spanish!!! – another possibility for presentation topic
Objective … of graphs, charts, diagrams, visualizations is to represent data. Typically data is sets of numbers or sets of other information. Examples: • height and weight of some set of people – possibly with other characteristics and over time • GDP, average longevity of countries – population, region, over time • frequency of words in a piece of text
Caution • If you don't have much data, then not much to do… • if data is wrong or definitions not understood or misunderstood, ….
Preview • Next Monday: guest speaker curator from Neuberger present artist visualizations
Organization / conference / competition / links • http: //www. visualizing. org/
2012 Olympic medal count • Issues – what counts? How to weigh types of medals? – Adjustment for population • what about wealth of country? • http: //www. nytimes. com/interactive/2012/0 8/07/sports/olympics/the-best-and-worstcountries-in-the-medalcount. html? ref=olympics
Gapminder • http: //www. gapminder. org/
Excel graphs • Construct data • Bar graphs • When are pie graphs appropriate? • x-y graphs • bubble graphs
Wordle • http: //www. wordle. net/
Dimension • The term 'dimension' is used for any attribute of a thing, not just height/width/length • Gapminder examples appeal (I think) because of quantity and dimensions of data – dynamic representation of time – color for regions – comparisons
Napoleon's March • … into Russia 1812 -1813 • Called by Edward Tufte (and others) the best (and still the best) graphical visualization. • Use Google to get multiple versions • Your opinion?
My Ds mnemonic • definitions • denominator • dimensions • [what's the] difference • data sources • distribution The point of visualizations is to show these things (and others? )
Examples • … and excuse to start reminding you to vote • http: //fivethirtyeight. blogs. nytimes. com/201 2/07/27/new-york-not-as-blue-as-it-couldbe/
Example • Drawn to show electoral votes (and assessment on how states will vote) – dimensions of size versus population http: //elections. nytimes. com/2012/electoralmap – Note: the geographic view does not accurately display Alaska or Hawaii in position and size!!! – Try multiple maps (scenarios) + our own?
If these visualizations make you less likely to vote because New York is predicted to be solidly Democratic… • popular vote will count in terms of mandate • there are down ticket contests • increasing the youth vote will make younger people be more important to politicians • good to get into the habit • people have died for the right to vote!
Example • Periodic table of visualization methods (!) Periodic table: http: //www. visualliteracy. org/periodic_table. h tml – assumes the periodic table has meaning for viewers
Timeline • http: //www. datavis. ca/milestones/ – on visualizations • http: //www. informationisbeautiful. net/visual izations/mountains-out-of-molehills/ – see also http: //www. informationisbeautiful. net
Tools • http: //www. computerworld. com/s/article/92 15504/22_free_tools_for_data_visualizatio n_and_analysis • http: //www. readwriteweb. com/archives/the _best_tools_for_visualization. php
Note / caution / repeat • If you don't have much data, then visualization does not help – if data is (potentially) misunderstood or misleading • Think about Ds
Homework • Decide on presentation topic and make proposal to the appropriate moodle forum • Postings: usual plus – any visualization topic • Tools • Edward Tufte: – Home page (not that easy to navigate and mainly tool to sell his workshops and books which are excellent: http: //www. edwardtufte. com/tufte/ » Forum: http: //www. edwardtufte. com/bboard/q-and-a? topic_id=1 » Critique of Power. Point for lectures! » etc. Next class: guest speaker on artist visualizations