Video summary
The video outlines a streamlined approach to conducting denominator analysis within the DHIS2 platform by leveraging imported external data sources. The core argument presented is that once trusted population data, such as census figures or UN projections, are successfully imported into the system, users can bypass the error-prone process of manually updating spreadsheets. Instead of relying on static Excel files, these external datasets become integral parts of the DHIS2 environment, allowing for dynamic comparisons between different estimation methods directly within the software's native tools.
To facilitate this analysis, the presenter demonstrates a specific dashboard designed to review denominators at subnational levels, such as provincial or district data. This localized focus supports a strategy of decentralization, where lower-level staff can verify their own denominator data using the "many eyes" principle to ensure quality without creating an overwhelming burden on a central team. By setting up validation rules and utilizing various data visualizations within DHIS2, the system automatically highlights discrepancies between different datasets, such as those derived from national censuses versus UN population projections, making it easy to identify where differences exceed a specific threshold, like 10%.
The practical application of this dashboard is illustrated through a detailed review of single-age population estimates. In the demonstration, the first column displays data calculated from census records, while the second column shows figures imported from UN population projection tools. The visual comparison reveals that despite the different methodologies used to generate these numbers, the values remain remarkably close for most entries, with only minor deviations noted in specific years. This high degree of alignment between independent data sources provides significant confidence in the accuracy of the denominators.
Ultimately, the video concludes that importing and comparing external data within DHIS2 transforms denominator management into a robust, automated process rather than a tedious administrative task. The ability to scroll through national totals and observe that differences remain consistently low, such as hovering around 5% between 2018 and 2019, validates the reliability of the imported estimates. This confidence allows analysts to proceed with greater assurance when using these denominators for subsequent data analysis, ensuring that both national and subnational estimates are built on a foundation of verified and cross-checked information.
Read the full video transcript
Okay, so what I want to do now basically
is walk through, you know, how some of
this analysis can be done in DHIS2.
Right? So,
one of the main things about doing this
analysis in DHIS2
it basically if you can import that
external data into your system then you
can start using it for compa risons.
Right? If I can take that population
data from the source that I trust or
from the source that I'm happy with to
use as my comparison. Um, you know, then
I don't have to keep updating a
spreadsheet over time, right? We all
know the the pitfalls of that. Though we
will use Excel a little bit just because
that's a a universal example we can use.
Um
you know, you can bring the data into
DHIS2 and compare it with what you have.
Right? And after you import that data,
you can DHIS2 tools, right? The
different data visualizations, um even
things like validation rules can be set
up um to compare these various
denominators. Um
and then we also suggest that you have a
dashboard that focuses on denominators,
right? Um particularly if you have uh
data at the subnational level that you
can review. So, provincial or district
data or, you know, lower level data
depending on your system.
Right? Uh because that can be reviewed
by lower levels, right? They can verify
their denominators and you could once
again decentralize part of this process
and we've talked about some of the
advantages of decentralizing the data
quality process in terms of, you know,
not making this
an overburdening exercise, um you know,
by splitting it up into small pieces,
having many people review the data, many
eyes on the data,
um and then it allows you to calculate
your national estimates or even
subnational estimates with more
confidence.
Okay, so
we do have a denominator dashboard
available in the training system.
Okay. So, if I just uh
log in, you can log in as your user,
right? Um there's a uh dashboard called
DQ {dash} denominators.
Okay. And it's going to have many of the
analyses that I discussed set up for you
to review.
In this first table, we have the
estimated single age population from the
census.
And then in the second column, we have
the estimated single age population
brought in from that uh UN uh population
projections, the the tool that I showed
you. We've imported that data, and then
we've made a comparison.
And basically, we've highlighted any
comparisons where there's a difference
of more than 10%. But for the most part,
if you look at this data, it is pretty
close, right? It It gives us a lot of
confidence here. Um given that we have
data that's calculated using a census,
and we're comparing it to data that's
calculated a completely different way
using UN estimates, and they're all all
within a very close range. And even
though there is some differences for the
single age target populations, um you
know, once I start adding this up,
you can see here at the bottom,
maybe it's a little small and I'll zoom
in. These are the national totals.
There's only a difference of 5% in 2018,
a 5% in 19. Okay, and I can keep
scrolling over, right? So, between 4 and
5%, right? This is pretty good, right?
This would give me a lot more confidence
in using these denominators subsequently
to analyze um you know, my data for each
of these years.