
Next to the Visual analytics page, the platform provides dedicated pages for specific spatio-temporal analytics. The first one is Trend analytics.
This page focuses on how things change over time.
Let’s click on Trend analytics in the navigation bar on the left. The page opens and loads some default statistics:
By default it focuses on the last day and the day before in the data set.
It compares the Average fare
between these two days and shows the trend during these days, the average, and the difference.
Including an indication of how this number increased or decreased.
It shows the trend statistics for the entire data set under Global Trend.
It also shows the statistics for local areas, in this case for the hexagonal cells. By default, the page selects 3 areas from your drawings and/or your uploaded Areas of interest GeoJSON files.
It uses a default UTC 0 time zone.
Let’s first focus on the global trend:
On the Global Trend card change the time zone to America/New_York - (UTC -4).
On the AREAS OF INTEREST panel:
You should now see that the global average taxi fare is around 20 USD with a decrease of almost 4% comparing the last with the previous day.
Let’s look at a single period of 24 weeks:
Under TIME RANGE
Click on '1'.
Select Last 24 weeks from the drop-down box.
The global average fare remains mostly constant with a small increase during the month of December.
Let’s identify when during the day the fares are highest:
Under PROPERTY click on By hour to group by hour.
You should now see that around 5am fares are highest.
Let’s bookmark our analysis, like we did before, naming it 'Average fare by hour of day'.
You can analyze local trends by filtering on areas. For this you can use the areas or shapes defined in your GeoJSON time series data. Or you can draw shapes on the map, or load additional area of interest GeoJSON files.
In this step of the tutorial we will draw two shapes ourselves: one surrounding Newark Airport and one surrounding JFK. We will then use these shapes for trend analysis to compare both areas.
Follow these steps to obtain the shapes:
Switch back to the visual analytics page.
Disable our data set layer Sample Data: Yellow Taxi New York
Search for John F. Kennedy
and hit Enter.
Zoom out a bit with the scroll wheel or by pinching.
Above the map select the circle icon and draw a circle on the map.
Give the shape the name JFK
and hit enter.
Now also create a circle shape surrounding Newark Liberty Airport:
Search for Newark Liberty
in the search box and hit enter.
Zoom out a bit to obtain an overview of the area.
Draw a circle shape and name it Newark
and hit enter.
Now return to the Trend analytics page and follow these steps to analyze and compare taxi fares in both areas:
In the AREAS OF INTEREST panel remove all areas and add our newly drawn JFK
and Newark
areas.
Select Average fare from the PROPERTY panel, choose None for Grouping.
Select two time periods and select Last Day vs Day Before from the TIME RANGE panel.
Collapse the Global Trend and scroll to the two cards with the Local Trends.
You can now see the Average fare change for the two areas we drew on the map and can compare. You should see that the average fare for pickups around JFK has decreased and is much lower than the ones at Newark.
Let’s bookmark this page and name it Last day fare change at JFK and Newark
.
All analyses on the xyzt.ai platform can be downloaded and saved to a .csv
file:
Right-click on the widget and select Save as CSV.
Or click on the 3 dots (…) in the top-right corner of the widget and select Save as CSV.
A CSV file with the statistics used to render the widget will be downloaded. You can then use this information for example in Excel, or for further data processing, for example using Python.
Go to the next part: Create a dashboard