A headline becomes a list of how often each word appears. “Training raised wages” and “wages raised training” can land on the same list, because order is thrown away. The first lab does this with two headlines at a time. The folder lab asks how many piles those lists should form. The number of folders is chosen on the slider. It is not written in the clippings.一则标题变成每个词出现了几次的清单。“Training raised wages”和“wages raised training”可以落在同一份清单上，因为词序被丢掉了。第一个实验一次做两则标题。文件夹实验问这些清单该分成几堆。文件夹的个数在滑块上选定。它没有写在剪报里。
Home首页 · Text as Data文本作为数据
Text as Data文本作为数据
Three labs start from the same short headlines about a training program. The first counts words. The second sorts nine clippings into folders. The third places words on a map by hand and scores the headlines again.三个实验从同一批关于培训项目的短标题出发。第一个数词。第二个把九行剪报放进文件夹。第三个把词亲手放到一张图上，再给标题打一次分。
A pitfall.一个陷阱。 The agreement score only sees the words you kept. Dropping little words such as the and for, or missing that two headlines use different words for the same claim, changes who looks alike. The score does not tell the press officer what the headline claims.一致分数只看见你留下的词。去掉 the、for 这类小词，或没看出两则标题用不同的词说同一件事，都会改变谁和谁像。分数不会告诉新闻官标题在主张什么。
Labs实验
Summarize, then compare先概括，再比较 starts with these word counts and ends on a match.从这些词数出发，走到一次匹配。
Concepts概念
Concept概念 Bag of words词袋 Concept概念 Document-term matrix文档—词矩阵 Put headlines in rows and words in columns. Each cell is a count. The word map is a different picture: each word is a place you chose, and a headline is the average of those places. The map lab compares that score with the count score.标题作行，词作列。每个格子是一个计数。词图是另一张图：每个词是你选定的位置，标题是这些位置的平均。词图实验把这个分数和计数分数比较。