Created
September 11, 2019 20:06
-
-
Save sgsinclair/4088164259c1f09967cb9bb66111c9be to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| <!DOCTYPE HTML><html><head> <meta charset='UTF-8'><title>Walkthrough 1: What is an electronic text and how can we study it?</title><meta name="title" content="<h1>Walkthrough 1: What is an electronic text and how can we study it?</h1>"><meta name="author" content="Geoffrey Rockwell & Stéfan Sinclair"><meta name="modified" content="2019-09-11T19:59:38.773Z"><meta name="created" content="2019-09-11T17:06:44.716Z"><meta name="language" content="English"><meta name="license" content="Creative Commons Attribution (CC BY)"><meta name="version" content="0.1"><style id="ace-chrome">.ace-chrome .ace_gutter {background: #ebebeb;color: #333;overflow : hidden;}.ace-chrome .ace_print-margin {width: 1px;background: #e8e8e8;}.ace-chrome {background-color: #FFFFFF;color: black;}.ace-chrome .ace_cursor {color: black;}.ace-chrome .ace_invisible {color: rgb(191, 191, 191);}.ace-chrome .ace_constant.ace_buildin {color: rgb(88, 72, 246);}.ace-chrome .ace_constant.ace_language {color: rgb(88, 92, 246);}.ace-chrome .ace_constant.ace_library {color: rgb(6, 150, 14);}.ace-chrome .ace_invalid {background-color: rgb(153, 0, 0);color: white;}.ace-chrome .ace_fold {}.ace-chrome .ace_support.ace_function {color: rgb(60, 76, 114);}.ace-chrome .ace_support.ace_constant {color: rgb(6, 150, 14);}.ace-chrome .ace_support.ace_type,.ace-chrome .ace_support.ace_class.ace-chrome .ace_support.ace_other {color: rgb(109, 121, 222);}.ace-chrome .ace_variable.ace_parameter {font-style:italic;color:#FD971F;}.ace-chrome .ace_keyword.ace_operator {color: rgb(104, 118, 135);}.ace-chrome .ace_comment {color: #236e24;}.ace-chrome .ace_comment.ace_doc {color: #236e24;}.ace-chrome .ace_comment.ace_doc.ace_tag {color: #236e24;}.ace-chrome .ace_constant.ace_numeric {color: rgb(0, 0, 205);}.ace-chrome .ace_variable {color: rgb(49, 132, 149);}.ace-chrome .ace_xml-pe {color: rgb(104, 104, 91);}.ace-chrome .ace_entity.ace_name.ace_function {color: #0000A2;}.ace-chrome .ace_heading {color: rgb(12, 7, 255);}.ace-chrome .ace_list {color:rgb(185, 6, 144);}.ace-chrome .ace_marker-layer .ace_selection {background: rgb(181, 213, 255);}.ace-chrome .ace_marker-layer .ace_step {background: rgb(252, 255, 0);}.ace-chrome .ace_marker-layer .ace_stack {background: rgb(164, 229, 101);}.ace-chrome .ace_marker-layer .ace_bracket {margin: -1px 0 0 -1px;border: 1px solid rgb(192, 192, 192);}.ace-chrome .ace_marker-layer .ace_active-line {background: rgba(0, 0, 0, 0.07);}.ace-chrome .ace_gutter-active-line {background-color : #dcdcdc;}.ace-chrome .ace_marker-layer .ace_selected-word {background: rgb(250, 250, 255);border: 1px solid rgb(200, 200, 250);}.ace-chrome .ace_storage,.ace-chrome .ace_keyword,.ace-chrome .ace_meta.ace_tag {color: rgb(147, 15, 128);}.ace-chrome .ace_string.ace_regex {color: rgb(255, 0, 0)}.ace-chrome .ace_string {color: #1A1AA6;}.ace-chrome .ace_entity.ace_other.ace_attribute-name {color: #994409;}.ace-chrome .ace_indent-guide {background: url("data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAACCAYAAACZgbYnAAAAE0lEQVQImWP4////f4bLly//BwAmVgd1/w11/gAAAABJRU5ErkJggg==") right repeat-y;}/*# sourceURL=ace/css/ace-chrome */</style><style id="voyant-notebooks-styles"> body.exported-notebook { font-family: helvetica, arial, verdana, sans-serif; font-size: 13px; font-weight: 300; line-height: 17px; -webkit-font-smoothing: antialiased } body.exported-notebook .notebook-code-editor { font-family: 'Monaco', 'Menlo', 'Ubuntu Mono', 'Consolas', 'source-code-pro', monospace; font-size: 12px; font-weight: normal; line-height: normal; white-space: pre; } .notebook-text-editor, .notebook-code-editor, .notebook-code-results { padding: 6px; margin: 6px; } .notebook-text-wrapper .x-panel-body { border: thin solid rgba(0,0,0,0) !important; } .notebook-code-wrapper .x-panel-body { border: thin solid rgba(0,0,0,0) !important; } .x-panel .notebook-editor-wrapper-hover.x-panel-body, .editable { border: thin dashed rgba(0,0,0,.2) !important; } .notebook-code-editor, .notebook-code-results { margin-left: 2em; } .notebook-code-editor { border-left: thin solid rgb(240, 240, 240); } .notebook-code-results { background-color: rgb(252, 252, 252); border: thin solid rgb(240, 240, 240); } .notebook-code-results .error { color: red; } .notebook-code-results pre { font-size: smaller; } .notebook-code-results .info { overflow: scroll } .notebook-code-editor-raw { /* used for raw code in saved view */ display: none; } .cke_button__sourcedialog_label { display: none !important; } .exported-notebook .notebookwrappercounter { float: right; } .exported-notebook .notebookwrappercounter a { text-decoration: none; } .notebookwrappercounter a { -webkit-border-radius: 3px; -moz-border-radius: 3px; border-radius: 3px; background-color: #f5f5f5; text-align: right; border: thin solid #d8d8d8; padding: 1px; margin: 1px; font-size: smaller; } table.spyral-table { border: thin solid #ccc; } table.spyral-table th { background-color: rgba(255, 255, 0, .05); } table.spyral-table td, table.spyral-table th { border-right: thin solid #eee; border-bottom: thin solid #eee; } table.spyral-table td:last-child { border-right: none; border-bottom: none; } .spyral-header, .spyral-footer { text-align: center; } .spyral-footer { margin-top: 1em; background-color: rgba(0,0,0,.01); border-top: rgba(0,0,0,.05); }</style><script> // this script checks to see if embedded tools seem to be availablewindow.addEventListener('load', function() {var hostnames = {}, warned = false;document.querySelectorAll('iframe').forEach(function(iframeEl) {let url = new URL(iframeEl.src);if (!(url.hostname in hostnames) && !warned) {hostnames[url.hostname] = true; // mark as fetchedfetch(url).catch(response => {warned = true;alert('This notebook seems to contain one ore more tools that may not be able to load. Possible reasons include a server no longer being accessible (especially if the notebook was generated from a local server), or because of security restrictions.'+url)})}})})</script></head><body class='exported-notebook'><header class='spyral-header'><h1>Walkthrough 1: What is an electronic text and how can we study it?</h1></header><article class='spyralArticle'><section id='g8hitelmgiplukmg' class='notebook-editor-wrapper notebooktexteditorwrapper'><div class='notebookwrappercounter'><a name="g8hitelmgiplukmg" href="#g8hitelmgiplukmg">1</a></div><div class='notebook-text-editor'><p align="center" style="text-align:center">What is an electronic text and how can we study it with a computer?</p><p>In this Walkthrough you will be introduced to the Voyant text analysis environment and the basics of text analysis. This will open questions about what is an electronic text such that we can use computer-based tools to study it. We therefore start with some general questions to which we will return at the end.</p><h3>General Discussion Questions</h3><ul><li>What is an electronic text?<ul style="list-style-type:circle"><li>How is it different from a printed text like a book?</li><li>How can we read differently with e-texts?</li><li>What is the interface of an e-text and how is different?</li></ul></li><li>How can we analyze electronic texts?<ul style="list-style-type:circle"><li>What can we study with electronic texts?</li><li>What questions can we ask with a computer that we wouldn’t ask by reading?</li><li>How can text analysis enhance or detract interpretation?</li></ul></li></ul><h2>Walkthrough: Word Clouds</h2><p>Let us look at a visualization produced by Voyant. This visualization, called <b>Cirrus </b>in Voyant, shows a cloud of words from a text. <i>Click the link to open the visualization and play with it.</i></p><p align="center" style="text-align:center">Word cloud of a text in Voyant<br><a data-cke-saved-href="https://bit.ly/2lH9One" href="https://bit.ly/2lH9One">https://bit.ly/2lH9One</a></p><p><i>Take a moment to reflect on this visualization.</i>Some questions to ask yourself:</p><ul><li>How is the computer selecting the words for us to see? Can you imagine the process by which the computer produces the cloud?</li><li>How do you interpret the visualization? What do the graphical features mean? For example, why are some words larger than others? What do the colours, orientation, and location in the cloud mean, if anything?</li><li>How can we control the visualization? What can we change and what would you want to change?</li><li>Can you guess what the text is? Does this summarize the text usefully?</li></ul><h3>Thinking-Through Word Clouds</h3><p>Word clouds nicely illustrate a number of aspects of text analysis that are so obvious they get forgotten.</p><ul><li>The visualization is the result of a process that starts with the <b>selection </b>of an electronic text, then <b>tokenizes </b>(splits) the text into words, <b>counts </b>the word tokens, and <b>generates </b>a coloured cloud of the high frequency tokens. This process was programmed as a tool by a developer who made a series of decisions that can affect the resulting visualization. When you use the tool you too are making decisions within the constraints set by the tool. Whose cloud is it at the end? Is it the text’s, the developer’s, or yours?</li><li>Some of the graphical features are <b>metrical</b>, in the sense that they are based on a measurement of the text, and some are not. For example, the size of the words is based on the frequency (how many times the word occurs). The colour has nothing to do with the text, but is randomly applied by the program from a colour palette (that you can select) to make the cloud look nice. Orientation is likewise random. The location in the cloud is loosely associated with frequency. Words near the centre tend to be higher frequency, but for visual purposes low frequency words are also fitted in. How else might you use the graphical features like colour? How else might you visualize the text?</li><li>The visualization shows some things and hides others. For example, it shows high frequency words, but discards word order. It treats the text as if it were a <b>bag of words </b>where all that matters is the count of words. The higher the frequency, the more important. What else is hidden? What else could be shown?</li><li><b>Visualizations </b>seem obvious until you think about how much you know about how they are generated. They seem obvious because we have learned to use certain genres of visualizations like bar charts, family trees, and word clouds. We see them in everyday situations, but need to be careful when using them as tools to interpret. How do they need to be tuned? How could the instrument mislead us? What are we missing?</li></ul><h2>Voyant Panels: Cirrus</h2><p>Let us learn a bit more about Voyant. Voyant has a collection of tools like Cirrus that show you the text though a panel. In this case we are seeing just one panel or one tool. Later we will see arrangements of interconnected panels. Let us learn about panels.</p><div align="center"><iframe src="https://voyant-tools.org/?view=cirrus&corpus=austen" style="width: 400px; height: 300px; margin-left: auto; margin-right: auto;"></iframe></div><p><br></p><p><b>Panel Controls: </b>Most panels have frequently used controls visible at the bottom. In the case of Cirrus we have a Terms slider that lets us control the number of words. <i>Try changing the number of terms visualized?</i>Do the words show up in the same place and colour as you change the number?</p><p><br></p><p><b>Help: </b>In the upper right hand corner there is a <b>? (Help) </b>button. (There are actually two – the top one is for Voyant generally, the one in the grey Cirrus area is for the panel/tool.) Mousing over gives you a short explanation of the tool. Click and you will get a link to the documentation with more detailed help. <i>Try clicking the link to the documentation and skimming the detailed help. </i>Can you get back to the visualization?</p><p><br></p><p><b>Options: </b>All panels have a roll-over area to the left of ? (Help) that gives you access to other panels and options. The Options button looks like this: <img data-cke-saved-src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAADsAAAA5CAYAAACI7VO4AAAMTGlDQ1BJQ0MgUHJvZmlsZQAASImVVwdYk9caPv/IJGEFIiAj7CXIEAggI4QVQUCm4CIkgYQRQ0IQcVtKFaxbRECtaFXEYqsVkDpR6yyC2zouSlGp1OLAhco9GVBrn3vvc7/nOf//5jvf934j5///cwDQq+PLZPmoPgAF0iJ5YlQYa2p6BovUA4hAH+gCe8DmCxQyTkJCLIAycv+7vLoOENX9ipuK65/z/1UMhCKFAAAkAeIsoUJQAPGPAOBlApm8CAAiG+pt5xTJVHg6xEZymCDEMhXO0eAyFc7S4Gq1TXIiF+K9AJBpfL48BwDdVqhnFQtyII/uTYg9pEKJFAA9MsTBAjFfCHE0xOMKCmarMLQDTlmf8OT8jTNrlJPPzxnFmlrUQg6XKGT5/Ln/Zzv+txTkK0diOMBBE8ujE1U1w77dzJsdo8I0iPulWXHxEBtC/EYiVNtDjFLFyugUjT1qLlBwYc8AE2IPIT88BmJziCOl+XGxWn1WtiSSBzFcIWiJpIiXrPVdJlJEJGk56+SzE+NHcLacy9H6NvHl6rgq+1PKvBSOlv+mWMQb4X9ZKk5O0+SMUYslqXEQ60LMVOQlxWhsMLtSMTduxEauTFTlbwdxgEgaFabhx2ZmyyMTtfbyAsVIvdgysYQXp8U1ReLkaC3PXgFfnb8JxK0iKSdlhEekmBo7UotQFB6hqR3rFElTtPVi3bKisESt73NZfoLWHqeK8qNUehuIzRXFSVpfPLgILkgNPx4nK0pI1uSJZ+XyJyVo8sFLQCzggnDAAko4ssBskAskHf0t/fCXZiYS8IEc5AARcNNqRjzS1DNSeE0CpeAPiERAMeoXpp4VgWKo/zCq1VzdQLZ6tljtkQceQlwAYkA+/K1Ue0lHo6WC36BG8o/oAphrPhyquX/qOFATq9UoR3hZeiOWxAhiODGaGEl0xs3wYDwQj4XXUDi8cDbuP5LtX/aEh4QuwgPCNUI34dYsyVL5Z/WwwGTQDSNEamvO+rRm3AGy+uBheBDkh9w4EzcDbvgEGImDh8DYPlDL1Wauqv5z7r/V8EnXtXYUDwpKGUMJpTh97qnrouszyqLq6acd0uSaNdpX7ujM5/G5n3RaCO8xn1tiy7AD2BnsBHYOO4y1ABZ2DGvFLmJHVHh0Ff2mXkUj0RLV+eRBHsk/4vG1MVWdVHg0evR5vNfMFYlKVO9HwJ0tmyuX5IiLWBz45hexeFKB+ziWl4enPwCq74jmNfWCqf4+IMzzf+kKjwPgXwGVOX/p+LYAHHoIAOPVXzrb5/DxWA3AkU6BUl6s0eGqCwFQgR58okyBJbAFTrAeL+ALAkEoiACTQDxIBulgJuyyGK5nOZgD5oMloBxUgtVgA6gBW8F2sBt8B/aDFnAYnAA/gwugE1wDt+Hq6QVPwAB4BYYQBCEhdISBmCJWiD3iinghbCQYiUBikUQkHclEchApokTmI18glchapAbZhjQgPyCHkBPIOaQLuYXcR/qQ58g7FENpqBFqgTqg41E2ykFj0GR0BpqDFqKlaBm6Eq1G69G9aDN6Ar2AXkO70SfoIAYwHYyJWWNuGBvjYvFYBpaNybGFWAVWhdVjTVgb/J+vYN1YP/YWJ+IMnIW7wRUcjafgArwQX4ivwGvw3Xgzfgq/gt/HB/CPBDrBnOBKCCDwCFMJOYQ5hHJCFWEn4SDhNHyaegmviEQik+hI9INPYzoxlziPuIK4mbiPeJzYRewhDpJIJFOSKymIFE/ik4pI5aRNpL2kY6TLpF7SG7IO2YrsRY4kZ5Cl5KXkKvIe8lHyZfIj8hBFn2JPCaDEU4SUuZRVlB2UNsolSi9liGpAdaQGUZOpudQl1GpqE/U09Q71hY6Ojo2Ov84UHYnOYp1qne91zurc13lLM6S50Li06TQlbSVtF+047RbtBZ1Od6CH0jPoRfSV9Ab6Sfo9+htdhq67Lk9XqLtIt1a3Wfey7lM9ip69Hkdvpl6pXpXeAb1Lev36FH0Hfa4+X3+hfq3+If0b+oMGDANPg3iDAoMVBnsMzhk8NiQZOhhGGAoNywy3G5407GFgDFsGlyFgfMHYwTjN6DUiGjka8YxyjSqNvjPqMBowNjSeYJxqXGJca3zEuJuJMR2YPGY+cxVzP/M6890YizGcMaIxy8c0jbk85rXJWJNQE5FJhck+k2sm70xZphGmeaZrTFtM75rhZi5mU8zmmG0xO23WP9ZobOBYwdiKsfvH/mqOmruYJ5rPM99uftF80MLSIspCZrHJ4qRFvyXTMtQy13K95VHLPiuGVbCVxGq91TGr31nGLA4rn1XNOsUasDa3jrZWWm+z7rAesnG0SbFZarPP5q4t1ZZtm2273rbddsDOym6y3Xy7Rrtf7Sn2bHux/Ub7M/avHRwd0hy+cmhxeOxo4shzLHVsdLzjRHcKcSp0qne66kx0ZjvnOW927nRBXXxcxC61LpdcUVdfV4nrZteucYRx/uOk4+rH3XCjuXHcit0a3e67M91j3Ze6t7g/HW83PmP8mvFnxn/08PHI99jhcdvT0HOS51LPNs/nXi5eAq9ar6vedO9I70Xerd7PJrhOEE3YMuGmD8Nnss9XPu0+H3z9fOW+Tb59fnZ+mX51fjfYRuwE9gr2WX+Cf5j/Iv/D/m8DfAOKAvYH/BnoFpgXuCfw8UTHiaKJOyb2BNkE8YO2BXUHs4Izg78J7g6xDuGH1Ic8CLUNFYbuDH3EcebkcvZynoZ5hMnDDoa95gZwF3CPh2PhUeEV4R0RhhEpETUR9yJtInMiGyMHonyi5kUdjyZEx0Svib7Bs+AJeA28gUl+kxZMOhVDi0mKqYl5EOsSK49tm4xOnjR53eQ7cfZx0riWeBDPi18XfzfBMaEw4acpxCkJU2qnPEz0TJyfeCaJkTQraU/Sq+Sw5FXJt1OcUpQp7al6qdNTG1Jfp4WnrU3rnjp+6oKpF9LN0iXprRmkjNSMnRmD0yKmbZjWO91nevn06zMcZ5TMODfTbGb+zCOz9GbxZx3IJGSmZe7JfM+P59fzB7N4WXVZAwKuYKPgiTBUuF7YJwoSrRU9yg7KXpv9OCcoZ11OnzhEXCXul3AlNZJnudG5W3Nf58Xn7cobzk/L31dALsgsOCQ1lOZJT822nF0yu0vmKiuXdRcGFG4oHJDHyHcqEMUMRWuREdywX1Q6Kb9U3i8OLq4tfjMndc6BEoMSacnFuS5zl899VBpZ+u08fJ5gXvt86/lL5t9fwFmwbSGyMGth+yLbRWWLehdHLd69hLokb8kvSz2Wrl368ou0L9rKLMoWl/V8GfVlY7luubz8xleBX21dhi+TLOtY7r180/KPFcKK85UelVWV71cIVpz/2vPr6q+HV2av7Fjlu2rLauJq6erra0LW7F5rsLZ0bc+6yeua17PWV6x/uWHWhnNVE6q2bqRuVG7sro6tbt1kt2n1pvc14pprtWG1++rM65bXvd4s3Hx5S+iWpq0WWyu3vvtG8s3NbVHbmusd6qu2E7cXb3+4I3XHmW/Z3zbsNNtZufPDLumu7t2Ju081+DU07DHfs6oRbVQ29u2dvrfzu/DvWpvcmrbtY+6r/B58r/z+9x8yf7i+P2Z/+wH2gaYf7X+sO8g4WNGMNM9tHmgRt3S3prd2HZp0qL0tsO3gT+4/7Tpsfbj2iPGRVUepR8uODh8rPTZ4XHa8/0TOiZ72We23T049efXUlFMdp2NOn/058ueTZzhnjp0NOnv4XMC5Q+fZ51su+F5ovuhz8eAvPr8c7PDtaL7kd6m107+zrWti19HLIZdPXAm/8vNV3tUL1+KudV1PuX7zxvQb3TeFNx/fyr/17NfiX4duL75DuFNxV/9u1T3ze/X/cv7Xvm7f7iP3w+9ffJD04HaPoOfJb4rf3veWPaQ/rHpk9ajhsdfjw32RfZ2/T/u994nsyVB/+R8Gf9Q9dXr645+hf14cmDrQ+0z+bPj5ihemL3a9nPCyfTBh8N6rgldDryvemL7Z/Zb99sy7tHePhua8J72v/uD8oe1jzMc7wwXDwzK+nK/eCmBwoNnZADzfBQA9He4dOgGgTtOc89SCaM6magT+E9acBdXiC8CuUABSFgMQC/coW+Cwh5gG76qtenIoQL29R4dWFNneXhouGjzxEN4MD7+wAIDUBsAH+fDw0Obh4Q87YLK3ADheqDlfqoQIzwbfuKtQZ+9DffCZ/BuUVn+XRvd8NAAAAAlwSFlzAAAWJQAAFiUBSVIk8AAAAgNpVFh0WE1MOmNvbS5hZG9iZS54bXAAAAAAADx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iIHg6eG1wdGs9IlhNUCBDb3JlIDUuNC4wIj4KICAgPHJkZjpSREYgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj4KICAgICAgPHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9IiIKICAgICAgICAgICAgeG1sbnM6ZXhpZj0iaHR0cDovL25zLmFkb2JlLmNvbS9leGlmLzEuMC8iCiAgICAgICAgICAgIHhtbG5zOnRpZmY9Imh0dHA6Ly9ucy5hZG9iZS5jb20vdGlmZi8xLjAvIj4KICAgICAgICAgPGV4aWY6UGl4ZWxZRGltZW5zaW9uPjY2PC9leGlmOlBpeGVsWURpbWVuc2lvbj4KICAgICAgICAgPGV4aWY6UGl4ZWxYRGltZW5zaW9uPjI0NjwvZXhpZjpQaXhlbFhEaW1lbnNpb24+CiAgICAgICAgIDx0aWZmOk9yaWVudGF0aW9uPjE8L3RpZmY6T3JpZW50YXRpb24+CiAgICAgIDwvcmRmOkRlc2NyaXB0aW9uPgogICA8L3JkZjpSREY+CjwveDp4bXBtZXRhPgq3yGnHAAAFEElEQVRoBe1Y309bdRT/FFpoKaUtLR3C3Ibb3C83UIMzQSPq00y2RBPdHs2yPeqDiW/GZ/8A45PG+SvxadEHEzQue9BtURmbziVjEkEZMFkLbWlpoT/wnMKlDfR77/devmSk3PNy23vP95zv53u+56cjnU4vYZtQ3TbBWYJpg61Va9uWtS1bAydgX+MaMGJVCLZlqx5LDby0LVsDRqwKwbZs1WOpgZdOFRj+m8thbHYR8UwB8WweucLGu8YmVx38nnq0epx4vM0Nb8PGL6FlsJlcEQPDSQyOp3EvsajizIQy6uscOECA+/Y0o6+rGQ4hp/4Hh5Xm/fb9DD75NYrYfF5f+iZ8PbTDg3PPhBH2mreTKbDZfBFf35zB5ZE55TDavC50d3rQ7nMh4K5HkTyB3WKCbs2NyXkks4VVnW664me6W9G/z2fKytJgR2cW8OGVaUTTaq2509+AN3tD2B92r4JZ+2OJgN8kwJ9dj2G24jYdISu/9VwEHgIvQ1JgF/JLeG9gAtOpnIxMKZ468sNXDvrx6hMBOOm3DHGc+GpoBj+Nlm/W810+nDsellkOKbBfDMXw492klEAZpmCTE2/3RfBYqFGGfR3P72Tlj649QJbAM73zQju6H/Gs41v7wtD+d6azuPRX+STXCrDy/zwFGKtAWV93RxNOk89q9CkFy/kV4Nq7ak9dsHx9PyZBS+w0iuhFCipH2o2tYKSuJId8lmk2k6frHTNaAl2wl0aSeKDQTzldnOkpW8RwdzoM7OVn6YZowenn0ZRhvtcFOzg+r6PO/KfXj7XC7dRVaUooH94JCnIaXb+nv1+hZs5xf1O6UUWuegd6KI+qpt5HvasiLYMdp2Su0lc5j6q0qoawo8WFEEV3Ji5b9cKL0LIJsqxKslLeyeoPrZSOBSq75hbE+xaCjVeUZ7JK9fhaqATcLPJXyNbbtxCsxylX1cgC0AoAWX4zfFlKkRp5dAKgEKyf+kiVpHfiG9UTpzyrEffAIhKCDa84vWih2fejMXWRvVI318tTNDxgYldpoKgvIiHY3cEGVPqCSIDse+59x2bUN/l/TGWQX5mMHDOoj4VgHXRAT3Y2yWKR4hsYTkjxyTJxmvmhokF5emc551aTIQTLzP17feBWTBVd+yeFoQn9KseMLj68kWi2tKSt2YWjBjW3Ltiu1kacOFAux8xsRMR74bcoUgvLrZmIR+b9VDKHi7fiJVYHXUMe1XCVpke6YHnha0cD6KBpgipKUP7+nCYOepWOka48FQ/cjS0Wlg/t5f0+HIyIJx2aPEOwPEXg/lPldf7l3xQ+uDxFHVU5ZWgbMnryeOj97ydXr2+Eru8b1GDIkNSkggVdvDWLb28vXxsZwTI8jVQAnO4J4qV9LYaDM7bmN3/G8d2dBIo8jSNiQ7zb3y5lVeaXBsvMV8dS+JJmQOlFcf3JfGZpb8iNZ3d7aQLhQYSmjJwJmLjWnSTf5KB2lYLbffqt0S5KjeePt2FXQN7FTIFlRdz6XRiM4obCqKoB4Ce7S0sjj1K5qC+u67x4YH7ycACnDvvBv82QabCa8CtkZZ4hV85ztW+b9dxD2eFsbxhc8Fghy2BZWYFOf3h6ga5ZulQdcf3LraEWJa1sSFvD45Yg1edBqnUP7XDjqU4vOv0u7bOl54bAWtL4EBcZpp6HuDflqm2wyo90iwi0LbtFDKF8G7ZllR/pFhFoW3aLGEL5NmzLKj/SLSJwW1n2f1RSq/0OYE3YAAAAAElFTkSuQmCC" src="data:image/png;base64,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" style="width:17px; height:9px">. Let us play with the options.</p><p align="center" style="text-align:center"><img data-cke-saved-src="data:image/png;base64,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" src="data:image/png;base64,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" style="width: 149px; height: 40px;"></p><p align="center" style="text-align:center">Hidden Panel Tools (Export, Change Tool, Options and Help)</p><p>The options for each panel will be different; lets look at the options for Cirrus:</p><p style="margin-left:36pt"><b>Stop-Words: </b>You can control what words are not displayed through a stop-word list. By default the stop-word list is turned on. <i>Turn it off (choose None) and see what words show up.</i>What sorts of words are in the stop-word list? Why are they “stopped”?</p><p style="margin-left:36pt"><b>Edit List: </b>You can edit and add words to the stop-word list. <i>Pick some of the words you don’t think are useful for interpreting the visualization and add them. <a>Try trimming words until you get a cloud that you think best represents the text.</a></i> How would you use this to make a poster for a talk?</p><p style="margin-left:36pt"><b>White (Word) List: </b>You can take the opposite approach and provide a list of the words you want to appear in the cloud. This is a white-word list. This will hide all other words. How can you use this?<i>Try creating a custom cloud with a white-word list?</i></p><p style="margin-left:36pt"><b>Colours: </b>You can also control the (colour) Palette used for the words. This is useful if you are trying to get a cloud that is attractive in some context, like a cloud for a poster. <i>Try experimenting with different palettes.</i></p><h2>Voyant: The Default Arrangement of Panels</h2><p>Voyant is not usually used one tool at a time. Typically we use “skins” or views of tools. These are arrangements of panels that are loosely interconnected. <i>Click on the link below to see the default skin.</i></p><div align="center"><iframe src="https://voyant-tools.org/?corpus=a60893516fca0d9034677c598d976df0" style="width: 90%; height: 450px; "></iframe></div><p align="center" style="text-align:center">Default View for Voyant: <a data-cke-saved-href="https://bit.ly/2kDlyH6" href="https://bit.ly/2kDlyH6">https://bit.ly/2kDlyH6</a></p><p align="center" style="text-align:center">This default view provides access to most of the commonly used tools. Starting in the upper left and moving around clockwise here are the default tools:</p><p style="margin-left:36pt"><b>Cirrus: </b>This is the word cloud tool that you have already tried.</p><p style="margin-left:36pt"><b>Reader:</b>This is a full-text reading tool so you can always return to close reading.</p><p style="margin-left:36pt"><b>Trends:</b>This shows you the relative frequency of top words over the segments of the text.</p><p style="margin-left:36pt"><b>Contexts:</b>This shows you a line of context for each instance of a word.</p><p style="margin-left:36pt"><b>Summary:</b>This provides summary statistics for the text you are looking at.</p><p><b>Try It: </b><i>Try playing with these panels.</i>Click on things and see what happens to the contents of other panels. There is obviously too much to play with in even the default view, but here are some hints as you try Voyant:</p><ul><li>Voyant is a research tool that is free so don’t expect it to be as robust as a commercial tool. Some tools are compute intensive and the server can slow down dramatically, especially if a whole class is experimenting at the same time. What can you do when things seem to freeze or you get lost? <i>Try Reloading the page – that will restart your session and bring you back to the place you started.</i></li><li>Learn how to use the help documentation. <i>Try loading the documentation in a different tab or window in your browser so you can refer to it as you go.</i></li><li>At the top of each panel to the left of the ? (Help) button there are often suggested other panels that you can flip in. For example, next to Cirrus you can try Terms and Links. All three of these tools allow you to see high frequency words, but in different ways. <i>Try these different tools. </i>What affordances do these different tools have? What are they good for and not so good for?</li></ul><p>Yes, there is a lot more to learn about using Voyant. Over different Worksheets we will learn to use specific tools and features. Yes, there are also other skins that combine different tools. Can you figure out how to switch to a different skin? We will also cover those further along. Now we need to see how to load a text into Voyant.</p><h2>Starting Voyant</h2><p>So far we have used links to combinations of panels and texts, but that is of no use to you if you want to use your own text. We will now learn how to use Voyant from scratch and load your own text.</p><div align="center"><iframe src="https://voyant-tools.org/" style="width: 900px; height: 550px; margin-left: auto; margin-right: auto;"></iframe></div><p align="center" style="text-align:center">Entry Screen: <a data-cke-saved-href="https://voyant-tools.org/" href="https://voyant-tools.org/">https://voyant-tools.org</a></p><p align="center" style="text-align:center"><br></p><p>If you go to Voyant (<a data-cke-saved-href="https://voyant-tools.org/" href="https://voyant-tools.org/">https://voyant-tools.org</a>) without specifying a corpus you will arrive at the entry screen.</p><p><i>Go to the Voyant Entry Screen. </i>The screen gives you some basic options on how to use Voyant and some ways of loading a corpus for analysis. Lets look at some of the options first.</p><p class="Feature">Language Interface Options: Voyant has different language interfaces like French, Arabic, and Japanese. These interfaces don’t affect the operation of Voyant so much as the labels and buttons. Voyant will try to detect the language of the text uploaded.</p><p class="Feature"><a>Global Options</a>: Voyant lets you set some global Options before choosing a text like how you want the text processed. We will learn more about these options later.</p><h3>Selecting and Loading a Text</h3><p>The single most important decision you will make when interpreting a text is selecting the text to study. There are a number of ways to do this:</p><p class="Features"><b>Paste in a text:</b><i>The simplest way to get your own text in is to just copy it from somewhere (a web page or Word document) and paste it into the box. Then click Reveal and the Default View will be generated.</i></p><p class="Features"><b>Paste in a URL:</b>You can also paste in a URL or a set of URLs (one per line) and Voyant will go to the page, copy the text, and use that for its text. If you paste in a set of URLs it will try to get each one and append them to each other to give you one long text.</p><p class="Features"><b>Upload:</b>Using the Upload button you can upload a text or a Zip file with multiple texts.</p><p class="Features"><b>Open:</b>The open button allows you to open an existing indexed corpus we have already loaded like Shakespeare’s plays or Austen’s novels. These corpora are here to make it easy to play with a collection when you are learning or showing Voyant.</p><p><b>Try It: </b><i>Try different ways of loading a text into Voyant:</i></p><ul><li><i>Try pasting in some text from a paper you are writing?</i>Can you learn anything about your own writing using text analysis?</li><li><i>Try finding and loading a text that interests you in the Gutenberg project web site. </i>Use the URL for the text and see what you get.</li><li><i>Try uploading a text.</i>Try downloading a text from Gutenberg, stripping out the Gutenberg introduction and rights statement (at the beginning and end) and then uploading the cleaned text.</li></ul><h2>The Electronic Text</h2><p>At the beginning of this Walkthrough we asked what an electronic text was such that we could study it. We have a form of answer already in the different texts we can load into Voyant, but there are other ways we could answer the question. To begin we can step back and catalogue some common types of electronic texts we use in everyday life.</p><p class="Features"><b>Born Digital Texts </b>are those that were created in a digital environment like email. They are distinguished by those e-texts that are attempts to represent a manuscript or print text.</p><p class="Features"><b>Word Processing Documents </b>are born digital or documents that are typed in. They are examples of e-texts created and manipulated by commercial tools often to be printed for use on paper. They typically have a lot of formatting and other information in addition to the “plain text.”</p><p class="Features"><b>Scholarly E-Texts </b>are typically editions of important analogue texts by scholars that are edited for study. They are often enriched with XML code (tags) following the Text Encoding Initiative. They come with metadata about the creation of the e-text so you know where it came from and what decisions were taken in the editing. These are meant to be rendered in some publication system so that they could be read on computers or over the internet. They could also be created for research purposes.</p><p class="Features"><b>Web Pages </b>in HTML (HyperText Markup Language) are the most common form of e-text we use. They are created not to be printed, but to be read over the internet. The markup language provides formatting instructions so that the author/web designer can control how the page looks. Web pages are often generated by server-side web applications from information in databases.</p><p class="Features"><b>PDF </b>(<a>Postscript Digital Format</a>) is a format for representing pages with text. Note that a PDF may be only a sequence of pictures of pages without any “text”.</p><h3>Excursus: Understanding Strings</h3><p><b>Strings</b>are how computers think of a text. A string is just a sequence of characters from a predetermined alphabet. Strings have no necessary information about all sorts of things we think are essential to texts like words, sentences, paragraphs, pages, meaning, language and so on. They are a data type that programming languages can manipulate. Understanding the gap between strings and texts is fundamental to understanding what computers can do with e-texts or not. A computer doesn’t read a text and it doesn’t understand meaning. All it can do is manipulate strings. Imagine that all you saw when you handled text were strings of encrypted characters like:</p><p align="center" style="text-align:center">Hello world! This is a string.</p><p>This is actually the phrase “Hello World! I am a string.” Even this example has meaning a computer doesn’t get in that it has spaces that you will automatically use to imagine words. For a computer a space is just another character that has no special meaning unless there are rendering instructions associated with it.</p><p>So how then can we analyze an electronic text? How can we use such a stupid system to interpret a text? One thing we can do is to break an e-text into smaller units like words. This is what analysis is – the breaking down of something into components to help you understand it and manipulate it. If we could define the boundaries of a component like a word in terms of characters then we could break a text into words and that would give us a <b>bag of words </b>that we could count. From that we could get a word cloud!</p><div style="border-style:none none solid; border-bottom-width:1pt; border-bottom-color:windowtext; padding:0cm 0cm 1pt"><p style="border:none; padding:0cm"><b><i>Stop Here</i></b></p></div><p><br></p><p><b>Demonstration: Tokenizing</b></p><p>A fundamental operation in most text analysis techniques is tokenizing. Tokenizing is the process of splitting a string into smaller strings called. Tokens are usually words, but they don’t have to be. You might want to split the text into paragraphs, or sentences, or even characters. <i>Try coming up with a reliable way to tokenize on words or sentences – how would you do it by hand?</i></p></div></section><section id='adkb0pabg48eitwl' class='notebook-editor-wrapper notebookcodeeditorwrapper'><div class='notebookwrappercounter'><a name="adkb0pabg48eitwl" href="#adkb0pabg48eitwl">2</a></div><div class='notebook-code-editor ace-chrome'><div class="ace_layer ace_text-layer" style="padding: 0px 4px;"><div class="ace_line" style="height:16.521739959716797px"><span class="ace_storage ace_type">let</span> <span class="ace_identifier">text</span> <span class="ace_keyword ace_operator">=</span> <span class="ace_string">"Here is an example string with different things that might make it hard to tokenize! "</span><span class="ace_keyword ace_operator">+</span></div><div class="ace_line" style="height:16.521739959716797px"><span class="ace_indent-guide"> </span><span class="ace_indent-guide"> </span> <span class="ace_string">"Numbers like 1000.00 can confuse things."</span></div><div class="ace_line" style="height:16.521739959716797px"><span class="ace_identifier">text</span><span class="ace_punctuation ace_operator">.</span><span class="ace_support ace_function">split</span><span class="ace_paren ace_lparen">(</span><span class="ace_string ace_regexp">/ </span><span class="ace_constant ace_language ace_delimiter">|</span><span class="ace_constant ace_language ace_escape">[</span><span class="ace_string ace_regexp ace_charachterclass">.!</span><span class="ace_constant ace_language ace_escape">]</span><span class="ace_string ace_regexp">/</span><span class="ace_paren ace_rparen">)</span></div></div></div><pre class='notebook-code-editor-raw editor-mode-javascript'>let text = "Here is an example string with different things that might make it hard to tokenize! "+ "Numbers like 1000.00 can confuse things."text.split(/ |[.!]/)</pre><div class='notebook-code-results'><div style='height: 30px'> <div class="info">["Here","is","an","example","string","with","different","things","that","might","make","it","hard","to","tokenize","","Numbers","like","1000","00","can","confuse","things",""]</div></div></div></section><section id='vjcp7s1t2khffcy3' class='notebook-editor-wrapper notebooktexteditorwrapper'><div class='notebookwrappercounter'><a name="vjcp7s1t2khffcy3" href="#vjcp7s1t2khffcy3">3</a></div><div class='notebook-text-editor'><p><span style="font-size:medium"><b>Try It: </b>Follow the link to the Tokenizing Notebook. This is a small program for tokenizing strings that you can edit to see the effects of different decisions taken by the developer. The program takes a string to tokenize and a set of characters to use to split the string. It works through the string and every time it finds one of the special characters it splits off a “word.” In the Notebook you can:</span><br></p><ul><li>Alter the text to try challenging strings. <i>Try typing in a string with contractions like “don’t” or numbers</i></li><li>Alter the list of special characters to split on. <i>Try different special characters?</i></li></ul><p><span style="font-size:medium">Can you adapt the program so that it will properly tokenize any text? What are the challenges?</span><br></p><h3>What’s Next?</h3><p><span style="font-size:medium">Tokenizing is a fundamental operation that gives us the bag of words. We can then do all sorts of things with the words like counting them and visualizing them. In the next chapter we will look more closely at what we can do with texts as bags of words. </span></p></div></section></article><footer class='spyral-footer'>© Geoffrey Rockwell & Stéfan Sinclair (Creative Commons Attribution (CC BY)). Created September 11, 2019.</footer></body></html> |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment