Data Format & Style
All the visualizations in Many Eyes accept the same data table format. (There's one exception: the tag cloud also understands free text.) A data table consists of rows where the values in each row are separated by tabs. The first row of the table should be "headers" that describe the columns. Here's an example table in the Many Eyes format:
| Fruit | Color | Calories |
|---|---|---|
| Apple | Red | 100 |
| Banana | Yellow | 120 |
The table has a header row (Fruit, Color, Calories) and two data rows. The first two columns are freeform text and the third column will be automatically interpreted as numbers. In each row the values are separated by "tab" characters, not spaces. This format can be easily exported from databases or spreadsheets. We have a size limit of 5 MB per file.
That's all there is to the basic format. Some visualizations add their own twist.
For example, in our map of the United States, text values such as "AZ" or "MA" will be
interpreted as state abbreviations. If you are interested in a
particular kind of visualization, read its explanation page before you upload your data.
You may also want to watch the step-by-step tutorial on how
to upload data to Many Eyes.
A last note: before you upload data to Many Eyes, you will have to register and sign in.
The Elements of Data Style
Half the fun of Many Eyes is letting others play with your data. When you publish your data, make sure to communicate what it means and why it is credible:Don't forget the units!
It may be obvious to you that your numbers are in thousands of dollars. But other people won't know "1,000" means a million bucks unless you tell them. Please label the units in all your data sets! If all your numbers measure the same quantity, include the units in the title ("2004 US Sales of Okra, $1000s"). If different columns represent different types of measurements, use the column headers instead.Adding more information
If a time or place is important, include it in the title. "Web Usage, Berkeley, 1993" is a very different data set from "Web Usage, Nigeria, 2007". We also have fields for source snd a source URL. When possible, enter the source of your data in the "source" line. If you have a URL where the data may be verified, add that as well. All of this information will help other people understand and trust your data. And if you have other notes, add them in the description field. Remember that you can comment on other people's data sets. We encourage you to ask for clarifications or point out features of the data that the original uploader may have missed.Punctuation counts
Currently the Many Eyes format accepts only two kinds of values: text and numbers. Commas, dollar signs, percentage signs, and the like are automatically removed from numbers. If you need to indicate numbers are percentages or currency, put that information in the column headers.
Common Problems & Considerations
Don't forget the header row
The header row is important: In addition to describing what the data actually means, the names are needed to set up and label the visualizations. Be sure your data has a header row! If you try this:
The first row will be interpreted as labels, not data.
Null or missing data points
If your data set contains some blank cells or a few "n/a" or "*" characters where numbers should be, that's not a problem--our graphs and charts will cheerfully adjust. One particularly tricky case, however, is when a column of data is almost entirely made of a markers such as "n/a". In this situation, our parser will guess that the column is text, not numbers. In the data upload page you can correct this by hand using the drop-down menu at the bottom of the column.
Rows should all have the same number of items
Each row should have the same number of items. (This includes the header row.) It's easy to copy only part of a data table by mistake, so if you see unexpected behavior it's worth double-checking the rows.
Multiple header rows
Some data tables contain several header rows, often with blank cells, to indicate multiple categories. For example:
Many Eyes currently doesn't handle multiple header rows, so in this case you would need to edit your data in a spreadsheet or word processor to merge the headers like this:
Summary rows
Many data tables contain summary rows. For example, the table below contains a row for "all citrus" that sums up all the other items. If you don't want a summary row to show up as a separate item in your visualization, delete it before uploading.
Rows versus columns
Sometimes you may want to switch rows and columns. For example, our line graph expects to graph the contents of columns. So if your data looks like the table on the left, you would need to switch rows and columns so it looks like the table on the right.
You'll usually get a chance to flip rows and columns when you configure your visualizations. (For technical reasons, if you mix more than one column of text with numbers we don't let you transpose; this may be changed in a future release.) If you know in advance you will have to transpose your matrix, it's also easy to do in Excel: Copy your data table, make a new worksheet, then choose Edit --> Paste-Special --> Transpose.