TimeFlow

is a tool created for reporters by Fernanda Viegas and Martin Wattenberg of Flowing Media Inc. that manages chronologies and timelines. It was envisioned as a way to: Keep notes on long-running stories, such as court cases, bankruptcies or police investigations that often require reminding readers and viewers of the events so far or the history of the case. Compile disparate material for projects in a way that might help see relationships and events in a new light. Organize information for narratives and reconstructions of events.

The key features include the ability to zoom in and out on any period of time, to filter and color code events, and to edit your notes in-place or import them from a spreadsheet. It wont automate your data entry, but it should help you organize and visualize the notes youre taking anyway.

Getting started
Take a look at the installation instructions on the TimeFlow github wiki at http://wiki.github.com/FlowingMedia/TimeFlo w/download-and-install-timeflow . You can install this into your own documents folder or onto a flash drive.

Looking at raw data

The Deepwater Horizon oil spill

To get started with TimeFlow, open the Example dataset called Deepwater Horizon Spill, accessed from the menu under Examples: Before digging in, take a look at what data is included in this timeline. Choose the List tab at the top of the main screen:

Your list will be sorted chronologically no matter what order they were entered in.

Getting oriented on the timeline

First, notice the dates. Unlike other tools, you can use approximate dates in TimeFlow -- it handles and displays dates entered as years, months, days or date and time. An entry with both a start and end date is shown as a span and a start date alone is represented as the event date: There are several types of data in TimeFlow. The key one is, of course, a date/time value to place the data on the timeline. A URL type links to source material; a list is a small set of values that might apply to a single record, such as the people or organizations involved in an event or subject tags used for colors and filtering. (Details are covered in the section on creating your own timeline) You can also view data in the timeline itself. Anytime you mouse over a point, you see all of the data for that point: This timeline defaults to the entire span of the data, including the 1990 law holding owners responsible for oil spills. The small section at the bottom is one of several ways to navigate through time. Click on a year, month or day to zoom in on that period. Because were looking at the BP oil spill, the first step might be to zoom in 2010. You might want to zoom in a little further to just months of April June 2010. You can do Switch back to the timeline view using the controls at the top of the screen. There are four sections to the timeline along with the main window:

this three ways: drag your mouse around the points you want to see in the large window

If you zoomed in too far, you can zoom out a little or all the way using the Zoom control. Here you can also choose how your timeline is displayed. Diagonal is good for small timelines like these, since it shows as much text as possible like a set of stairs:

...or narrow the band on the small navigation pane on the bottom of the timeline by clicking and dragging its edge: Finally, you can double-click in an area of the main timeline window to zoom in on that period. (It doesnt work well until you start narrowing in on a period of interest.) You can scroll through your timeline by dragging the yellow area in the navigation pane at the bottom.

For larger timelines, this view sometimes makes you think a pattern exists when it doesnt. Then, a loose view might be better:

Controls in a timeline
There are two areas of controls in the timeline view. In one, you control how the view is displayed. In the other, you control the labels, colors and groupings. Look for the most basic controls in the upper left panel. It changes with each view of your data (calendar, timeline or graph): For a very large timeline, you might use the graph view to help you navigate to areas with a lot of entries. The lower left section of the controls stay the same no matter what view you have of your data:

the first point in your timeline, in this case the 1990 law holding oil companies responsible for spills. Scroll down to a more interesting time:

(Dot size is reserved for data that has number values in it, such as campaign contributions. In this case, if we had the size of the oil slick, we might choose to use it.) Try playing around with it, envisioning various stories. In this view, all of the groups have been removed and weve selected the red BP actor to see, in context, what BP had done during the spill: The calendar keeps the settings you made on the timeline, so it is highlighting the BP actions. Days in which there are no events are grayed out. Notice how the options on the upper left of your screen have changed. For large timelines, this view of the dots will show you the most. But for small ones like this, you might want to see the labels.

All of the points are still visible, but the ones not selected are grayed out to highlight your story.

Sometimes youll see half a dot, or a dot with more than one color. That means its got more than one category chosen:

Calendar view
Sometimes looking at a calendar is easier than a large timeline. When you click on Calendar on the top navigation menu, it always takes you to

This timeline opens up to a slice of the first 100 days, visible on the bottom navigation view:

Practice questions for viewing timelines and calendars:

1. When did President Obama first visit the spill site? Where was he the day before his visit? 1 2. Did BP executives see appear before Congress before or after oil began washing ashore? 3. When did the Interior Departments Minerals Management Service, or MMS, propose stricter safety rules on offshore rigs? (In each of these questions, zooming in and out, selecting different actors to highlight in color and switching between calendar and timeline views will help you see patterns easily.)

Zoom out to 100% and take a look at the global controls on the left to get yourself oriented. You may not have noticed it, but there is another indicator on the left menu:

This tells you that all of the items in the dataset are being shown there is no active filter. Right now, the data is being shown by the place of the event the White House, D.C. or somewhere else. The most frequent places are listed first on the list. To get a different view of the data, try switching to a view that shows tags with no groupings, then highlight stimulus:

Filtering
Once a timeline grows to more than a few dozen events, there is simply too much to digest in one sitting. That is when the tools filtering options come into play. For this example, well use a different example dataset: the Obama-First 100 Days:
1

Note: In early releases of the software, there was an error in the sample data showing Obama in the Gulf Coast on Apr. 28. That date is really May 28, which was corrected in later releases.

Now you have a list of items you can choose from in the left hand column: The color map reminds you how to read your timeline. While there is still way too much to look at, it becomes easier to see the patterns of the stimulus bill, which he signed on Feb. 17. Switch to the Filter view using the tab on the top left: These tags were made up as the data was being typed in. They are not mutually exclusive an entry could both be a reversal from previous policy AND an energy action. Lets first look at how quickly Obama reversed policies of the Bush administration, and what those first actions were:

At first, all you see is an empty search box. But you can add in any of your categories from the dataset under the menu item Filters: (If your example looks different, flip back to the display view and switch it to Diagonal on the upper left panel.) i Add tags, who and entry type and location to the filter list by checking them off on this menu: (You may still have a color highlight shown. If you do, flip back to the timeline and clear all of those selections by clicking on them again.) Its pretty simple now to look through the events that marked a reversal. Now try looking at the stimulus program:

This is much bigger and might need more filtering. Lets look just at trips taken to promote the stimulus:

Editing a timeline
There are several ways to edit your timeline. The easiest way is on the timeline itself. To follow this example, clear all of your filters on the Obama first 100 days example from either the Display or Filter view:

(Note: the counts you see next to each item are unfiltered counts. In the example above, there are 22 entries for all trips, not just those for the stimulus.) This narrows your view to just nine items:

Ive also turned off the grouping and colors to get a clean view of the timeline, and narrowed my view to late January: To edit, right-click on a point in the timeline:

You can see that the better you tag your data, the easier it will be to use later on. Finally, you can do the equivalent of a full-text search for keywords or phrases:

Regular expressions (not case sensitive) are also available, but not documented, in the search box. In this case, it searches for Michelle, Sasha or Malia.

You can delete the event, edit it, or add another on the same day. (If you didnt select an existing

event, youd just have the opportunity to add one.) If you choose Edit, youll get a form to fill out:

Add fields the kinds of information you want to enter, such as a label for the timeline, a long description or tags, under Edit, Add field. Add records an entry in the timeline in the same way, or directly on the timeline or calendar itself.

Importing
Right now, importing from a spreadsheet replaces all of the existing data there isnt a way currently to add to your dataset. Importing from a spreadsheet is done two ways: by copying and pasting, or by saving as a text (TSV or CSV) file. Most often, youll just copy and paste. Adding a lot of events is probably easier from the calendar view: One trick in using spreadsheets as your data entry is to define your date fields as text rather than General types. The reason is that TimeFlow is more forgiving than Excel in its date formats and the precision it expects, so it can handle information that cant be well sorted or represented on a spreadsheet: The advantage of using these views over the List and Table views is that TimeFlow keeps your place as you edit. On the other views, you will have to scroll back to where you left off.

In this case, we might want to add people to this National Prayer service. (If youve defined your data field as a list, the way this one is, separate items with a comma.)

Creating a timeline
If you havent already typed your information into a spreadsheet, one way to create a timeline is to start from scratch. In the menu, choose File, New from Template, to choose some existing setups. You will probably use the template Events with start and end dates as your model.

To get the data into TimeFlow, select all of your cells and copy. This data will replace whatever you already have open, so consider starting a new timeline to make sure you dont lose any other work.

 Do the same thing with the tags field. Select the text in the import box, then paste: This data also has a column for the direct link to stories that acted as the source for the information. TimeFlow correctly guessed it was a URL field, but it might not have.

Best practices
There are also a few best practices for reporting with timelines built into this data. First, there is a source for each row. This dataset was built from news reports, so it simply records which news source and then the link to the story. There is also a field called work status, which says whether or not the data has been fact- checked enough for publication. That way, you can filter for only information you need to check, or alternatively only information you feel comfortable publishing. It also has an ID row number, which helps when you want to sort in the order you entered the data. For instance, you may have accumulated many points from a single document spread over a long period of time. This will help you get back to the way you saw it originally if you have to fact-check. Well want to change those guesses in some of these cases. The actors are used as tags for filtering each row has a list of actors in the event, and we want them to be split up when they are imported. Change text to list when you want TimeFlow to treat it that way: Once youre happy with your decisions, click the import this button. You may still have to check some default settings. In this case, the data had fields that were named start and end, which TimeFlow recognizes as the key date fields in your data. These are the ones that will be used to place the points on the timeline. But you might have

Once you import, TimeFlow will guess what kinds of fields youve imported and give you some basic statistics about them:

other names, and you might have multiple date fields. The fields used for your timeline are determined under Edit, Set date fields menu:

In this case, we need to add the end date to use as the ending period of a span of time:

In this case, TimeFlow didnt understand the Sept abbreviation for Sept. 14, 2009. You can just fix it in the List view: Once its fixed, you can erase the entry in the unparsed fields so that you only see the ones that are still problems. You can check to make sure none of the dates are missing by checking the Summary tab at the top right:

Notice, too, that there is at least one record that didnt have a good date: it shows that 97 percent of the start dates were defined. This is a good way to check to make sure your data was complete: One of the easiest ways to check your data is to go to the filters, then add the new, Unparsed fields field to your list.

You can select all of the non-missing UNPARSED FIELDS to look at the bad data on the list view:

Exporting
Right now, there is no way to print a timeline or its entries. To get a simple list, use the File, Export HTML to get a full listing that can be printed in Word or another text editor. Export a TSV to import into Excel or another spreadsheet program. (This stands for tab- separated values, and will be imported as a text file in your spreadsheet or database program.)

FAQ
Q: My timeline doesnt show up. The dates seem to all be there, but the entries are in 1969. A: Check the menu item Set date fields to make sure it understands which columns you want to use as your dates. It should guess, but sometimes it doesnt. Q: Not all of my data got imported. There are several rows missing. A: One type of entry is particularly difficult to parse: those with long narratives that end with a quotation mark. Go back to your original data and add a space after the end of the entry. Q: My list fields didnt separate out. They are still being shown as one long piece of text. A: Check your delimiter. It will separate based on commas.