Question 1

How do I create a comparative alert which compares indicators?

Accepted Answer

Navigate to the Alerts feature and click on 'Create Alert.'

Select the type of alert you want to create, i.e. Threshold alert or Comparative alert.

Threshold alerts allow you to compare indicator values to a set value, e.g. if you set an alert for Total malaria positive cases greater or equal to 100, the alert will get triggered when the positive malaria cases are equal or greater than 100.
Comparative alerts allow you to compare two indicators. These can be used to alert users in the case of data quality violations, e.g. when comparing the number of Malaria tests conducted to the number of Malaria positive cases. The alert will get triggered when the number of malaria tests conducted is less than the number of malaria positive cases.

Select the first indicator for your comparative expression.
Select the comparison statement to be used for your alert, e.g equal to, greater than, etc.
Select the second indicator for your comparative expression.
Select the ’group by’ you want your data to be disaggregated by.
Select a filter to be used on your ‘group by’ if applicable.
Select the frequency by which the alert should be evaluated.

Question 2

How do I select a time period (“Date Range”) filter for my query?

Accepted Answer

An additional possibility to restrict the scope of your analysis is to filter the data by time period. Adding a date filter allows you to pull data on all indicators from the time period you specify in the filter.

To select a time period for your analysis:

Go to the Analyze tool page.
Under the ‘Filters’ section, click on ‘Add+’ and then on ‘Date Range.’ This will add a date tag in your query form to customize.
Here you can choose either a custom date range or from a number of preset options including ‘this’ or ‘last’ week, month, quarter, or year as well as ‘all time’ and ‘Year to date.’ If you select “Custom” in the date range dropdown, then you will be able to filter to specific start and end dates by using the input calendars. You can also query ‘since’ a specific date.

Note: If you don’t apply a time filter to your analysis, by default all data available (‘all time’) in the platform will be added to your analysis. Also note that the date selections are “inclusive”; this means the days you include in the range will also be queried.

Question 3

How do I set a goal line for my visualization?

Accepted Answer

A goal line is a target, threshold or reference line you would like to make visible on your bar chart. Having a goal line on a chart allows you to show if the various bars are over or under this threshold.

To add a goal line, go to ‘Axes’ tab in ‘Settings'. You can add a horizontal goal line by setting a value and label. Note that in order to set a goal line for a percentage, you must provide the goal value in decimal format (e.g. 95% should be entered as 0.95).

*Note all data displayed in this manual is publicly available and is not confidential in any way or form.

Question 4

How do I apply different query operations other than ‘SUM’ to my analysis?

Accepted Answer

To apply operations other than the ‘Sum’ function to your indicator:

Click on the indicator tag in the query form, and
On the dropdown menu under ‘Operation’, make your selection.

The new operation will automatically be applied to your selected indicator.

Note: If you would like to compare different operations side by side, you can select the same indicator more than once and apply a different operation to each duplicate indicator in your query.

Question 5

What is the Data Outlier Analysis tab in Data Quality Lab (DQL)?

Accepted Answer

This tab aims to help you identify potential data quality issues based on outlier data points.

The score is calculated based on the proportion of data points that are moderate (2-3 standard deviations from the mean) and extreme outliers (3+ standard deviations from the mean). The higher proportion of outliers, the worse the score will be.

Below this score explanation is a box plot which you can adjust using the geography filters, aggregation and time filters.

The box plot surfaces facilities based on the proportion of their data points that are outliers so you can quickly identify the most problematic facilities.

Clicking on a dot in the box plot will surface the reported values from that facility in a time series visualization for you to assess the outlier data points to see if there’s an issue.

You have the option to switch the box plot to a table visualization where the rows are facilities and columns are the % outliers - this way you can see all facilities and isolate specific ones of interest (vs. the box plot doing a good job bubbling up outliers but not allowing you to do a comprehensive or specific check)

*Note all data displayed in this manual is publicly available and is not confidential in any way or form

Question 6

What is the Box Plot visualisation?

Accepted Answer

The box plot enables you to understand the distribution of your data (i.e. min, max, median, quartiles etc.), highlight outlier data points for investigation and the density of the data at different values using the violin plot. The boxplot can be used across a wide range of use cases including identifying outliers for epidemic investigations, reporting data quality issues and understanding seasonality of disease burden over time.

To successfully query a box plot, it is very important to use more than one group by and then use the settings tab to ensure you select the correct disaggregation under the setting “Display grouping level”. Try reordering your group bys to change which group by is used to be your “boxes”.

The tooltip of the box plot now shows you the label for all group bys when you hover on a dot. This means if you've grouped by province and each dot is a facility, you will be able to see the facility name.

*Note all data displayed in this manual is publicly available and is not confidential in any way or form.

Frequently Read Articles

Browse All Categories