We've disabled the flower visualization while we work on making it more useful for exploring your data. Check back in future versions for a new, improved flower!

The correlation visualization is disabled because there are fewer than two streams visible.

To enable the correlation visualization, make at least two streams visible.

Converts each data point to a rolling average (an average of the previous data points). Good for seeing underlying trends in noisy data.
The rows tool divides up your dataset according to its value – high, medium, or low, etc. From here you can change the colors or thresholds that are used for the Flower View .
The time shift tool visualizes your data with a time offset.
The Y-shift tool adjusts how the y-axis is displayed. It does not change any values. Dragging the arrows will stretch and shrink your data. Dragging the middle of the bar moves the whole line up and down.
Lets you zoom using your scroll wheel to a particular data point. Point your cursor to the data of interest and scroll. Clicking in the space between the handle bars will fully zoom you out.
The zoom you choose applies across all visualizations. That means that if you are only viewing two months on the time series, and click over to periodic pattern , only two months’ worth of data will be included.
Chooses how far back in time to go to include previous data points in the rolling average.

Quantiles

Divides up your data into equal parts. For example, if I had 100 data points, and I selected quartiles, the lowest 25 data points would be in the first quartile, the next lowest 25 in the second, etc.

Equal Ranges

Divides up the Y-axis (where the values are) into equal parts. If your lowest data point was a 1, and the highest 100, the first row would be 1-25, the second 25-50, etc.

Whiskers show (from bottom to top) the minimum, third quartile, median, first quartile, and maximum for data in that time period.
Absolute scaling is currently turned on. Data in each time period is scaled relative to data from the other streams, across all time periods.
Absolute scaling is currently turned off. Data in each time period is scaled relative to data from the stream itself, across all time periods.
See and search annotations.
Exported data will match what is displayed in the Timeseries visualization, and will reflect changes that were applied there. These changes may include rolling average, offsets, and so on.
These are the rules that Data Sense uses to create the Periodic Pattern view and Flower view . By default, we show the average of each data point that occurs in that time slot (bin).
To make your own workweeks vs weekends, commute hours, etc. click “New” in the Time Bins dropdown. Each time slot (Monday, Tuesday, etc.) is a “bin”.
Changes the value Data Sense uses to determine high medium or low, etc. Slide the black dot up or down to your desired level. From here you can also control what color is used for each level. For more precision, use the rows tool in Time Series.
Menu lets you change the times shown. Clicking on “Time 1” displays the labels directly on the flower.
Menu lets you change the times shown. Clicking on “Time 2” displays the labels directly on the flower.
Shows data as map pins. In this mode, one or more streams can be shown at once. Black pins represent locations for which data from more than one stream is visible.
Shows data as circular orbs. In this mode, only one stream is shown at once, and the orb size reflects the values for that stream at that location.
Draws a box around data to create a new subset.
+1 means this always goes up with that, -1 means this always goes down when that goes up.

Lets you change how your data is grouped before being correlated.

In general, you are more likely to get a better indication of whether there is a correlation if you correlate by hour or day than by week or month.

Data Sense has to group data together when sample rates are not the same. If steps are collected every minute, but mood is collected once a day, then it can only correlate by day.

Some data, like steps, should be summed to make a total for the day, while others, like mood, should be averaged across the day. Only you really know what is most appropriate for your data.

In general, you are more likely to get a better indication of whether there is a correlation if you correlate by hour or day than by week or month.

Correlations could not be computed for the currently selected timespan.

To see correlations, select a different timespan using the “Correlate by” selector below, and/or change the correlation settings for the streams in this experiment using each stream’s menu.

Making your data available for anonymous aggregation pools data with other users, allowing others (and you) to make anonymous queries. Any time your data is queried, Data Sense gives you back the results, so you can see how others are using your data. Learn how we protect your privacy in this feature here.
Create social experiments by aggregating data that other users have shared.
Find Data Builder templates that you can use with your data.
Drag your sources to the area above to work with them. You can use as many sources as you want, then whittle and crunch them with the other operators.
Include or exclude data, based on some criteria. Start with Filter Stream below, then connect a filter at top and a stream to be filtered at left.
Crunch your data by transforming and combining it to answer your questions. Use Line Up Data or Time Shift to help line up data that isn't at exactly the same time.
Output your data for use elsewhere in Data Sense.
Run a calculation over many sources of data at once.
Whittles a specified stream down to only the data included by a specified filter.
Whittles two streams down using a two-stream filter. For instance, selecting an area in the correlations visualization makes a two-stream filter, one that includes data in the selected area.
Defines how filters are put together.
Includes data only if it meets criteria in every filter attached.
Includes data only if it meets criteria in any filter attached.
Includes data only if it does not meet the attached filter's criteria.
Includes data based on time.
Includes data before or after a specified date.
Includes data during "bins" (weekends, commute hours, etc.) you previously set in an experiment as part of a time cluster.
Includes data during a specified time range.
Includes data based on selections you previously made in an experiment. To use these, make selections in your experiments, then click "View in Builder".
A selection of one or more time bins from the periodic patterns visualization.
A selection of one or more flower parts from the flower visualization.
An area selection from the map visualization.
An area selection from the correlations visualization.
Includes data based on values.
Includes data higher or lower than a specified value.
Includes data with the specified value. You can use this with both numeric and textual data.
Includes data in a specified value range.
Includes data that matches a specified value or set of values.
Apply arithmetic (+ - × ÷) on streams and/or numbers. Arithmetic on two or more streams is applied to the values that line up in time. If much of your data did not occur at exactly the same time, use Line Up Data to make them both hourly, daily, etc.
Adds data from two or more sources.
Subtracts the data connected at the bottom from the data connected at the top.
Multiplies data from two or more sources.
Divides the data connected at the top by the data connected at the bottom.
Adds a number to each point in the specified stream.
Subtracts a number from each point in the specified stream.
Multiplies each point in the specified stream by a number.
Divides each point in the specified stream by a number.
Converts numeric values into their negatives. For instance, 2 becomes -2, -4 becomes 4, and so on.
Converts numeric values into their reciprocals. For instance, 2 becomes ½, 4 becomes ¼, and so on. For streams that have units, this also "flips" the units.
Inserts data in gaps where there is no data.
Based on the location data connected at the bottom, guesses your location at the times in your non-location data connected at the top.
Merges two or more streams together when the data types and units are compatible. Click on the box while in the work area to choose how overlapping data is handled.
Converts raw data into a scale centered at zero, based on how far each value is from the average. Useful for comparing highly variable data (e.g. calories) with flatter data (e.g. skin temperature).
Converts each data point to an average of the previous data points.
Compute statistics (min, max, average, etc.) on your streams. To see the result, use Display.
Compute the minimum value of the stream.
Compute the average value of the stream.
Compute the median value of the stream. This is the value such that half the points are higher and half are lower.
Compute the maximum value of the stream.
Compute the given percentile on the stream.
Compute several quantiles on the stream. For instance, 10 quantiles gives you the 0th, 10th, … 90th, 100th percentiles.
Compute the number of points in the stream.
Compute the standard deviation of the stream.
Provides operators for working with textual values.
Converts textual values into a count of how many characters it contains.
Converts textual values into a count of how many words it contains.
Provides operators for working with time.
Line up data by grouping it into coarser buckets (e.g. minute-by-minute data becomes hourly or daily readings).
For a given time range and fill type (e.g. hours, days, weeks, months), inserts data in gaps where there is no data.
Moves data backwards or forwards in time. Useful for yesterday vs. today or last week vs. this week comparisons.
Selects data within one data stream by when data is present in another data stream. Click TODO: icon to learn more about why this is useful.
Selects data within one stream by when there is data present in another stream, allowing for inconsistencies in time stamps. About When will only output the data stream connected at the top.
Selects data within one data stream by when data is absent from another data stream.
Provides operators for dealing with units.
Converts data to a specified unit.
Converts the data connected at the top to the units of the data connected at the bottom.