Link Search Menu Expand Document
Mobile Data Collection toolbox

9.1 Choose your tool


Multiples tools exists to analyze and vizualize data collected through MDC. To know what tool(s) is(are) the most adapted to your needs and resources can be complicated, all the more so as it might require more than one tool or approach based on various criteria.

Note: Keep in mind that -although presented here quite late in the process to be coherent with the data management cycle- you should be deciding on the tools for your analysis at a very early preparation stage, to ensure that the way you design your data collection will be coherent with the wished-for analysis.

We will try to give you some advice on different aspects, such as the type of analysis you wish to make, the type data you have to analyse, the technical or technological environment you are in as well as other criteria such as the human component. We hope this checklist will help you choose the best tools for the various analysis you will be making of your MDC data!

TABLE OF CONTENTS


What are the types of analysis you wish to produce?

The type of analysis you want to make (how far you wish to go on the statistical front, also if you are analysing this data just this once or on a regular basis…) will have a big impact on the type of tool(s) you might want choose. Here are different considerations you may want to look into:

  • Do you need thematic or sector specific analysis? Before delving into any type of advanced statistical analysis, think of how sustainable you need your organization to be on this analysis and how it should be catered into your choice of tool. Sometimes, advanced tools may take you much further for a given analysis but be a barrier for the replicability of the analysis at a future moment when the team in place may no longer be the same with the same skillsets. Taking this in consideration, there are numerous advanced tools for statistical analysis, such as R, SPSS, Stata or Sphinx depending on your needs. Your sector of intervention may also have dedicated analysis tools that it can be worthwhile exploring, like ENA for nutrition assessments in emergency contexts for example, that can already be to a certain extent set up for your needs, or more widely dedicated tools for the whole data chain of your specific process that would be more relevant than generic tools.
  • Is it a one-shot survey? For one-shot survey analysis the tool that is most often used in the sector is by far Excel, with features to process your data further before building graphs or pivot tables if you want to rapidly calculate sums, percentages, etc. You can however go further with tools such as those mentioned above.
  • Is it a regular data collection you are making? (ie. midline or endline of the same survey? Collection of data done in a permanent way on the maintenance of infrastructure?) If you need to run regular follow-ups of the same topic, then you can setup an Excel “table” mode & dashboard, in order to be able to import new data with all your analysis already set up. Similar processes can also be set up with some of the advanced analysis tools or sector specific tools listed previously (and can sometimes do it better, but at a higher budget or requiring more skilled human resources).
  • Do you need to link information from different databases or datasets? In that case a good solution is to use the offline version of Power BI (or another business intelligence tool such as Qlik or Tableau), which makes this process relatively straightforward. If you do not wish to use Power BI and are at ease with Excel, you can use functions such as “vlookup” to link two datasets together- it will be a bit homemade but can often do the trick for databases that are not too heavy. These solutions only work however if you have set up a unique ID in your datasets!

What type of data do you have to analyse?

The type of data you have collected may be a differentiating factor in terms of the tool(s) you may choose to analyse your data.

  • If your have pictures: Pictures can be treated in different tools depending how you want to display them. The easiest is often to look at them directly through the visualization features of your MDC tools (on the map, in a photo gallery, by accessing it through the table view), but you can often find a way to integrate them in whatever analysis tool you have decided to use for other purposes (uMap, Power BI etc). Depending on the type of usage you want to make of these photos, it is also possible to import them in Excel or Word by using an advanced procedure such as this one .)
  • If you have multiple choice questions: Multiple choice question output data is displayed in a way hardly exploitable as such because every option creates a new answer column. In Excel, it therefore requires a few extra steps, where you will count occurrences concerning the data in question to build graphs or pivot tables on them. For a tool that does it more naturally, you can see these Humanitarian Data Solutions video tutorials that explain how to do it with Qlick.
  • If you have repeated data: Loop data is data that comes from a series of questions imbedded in a survey that can be asked multiple times (for example, if you ask similar questions to all the members of a family). It is technically challenging data to analyse as they correspond to a different database than your main database. Thus th particularity of the output data is that when you download them in Excel format all the content of answers from the loop will appear in another sheet. If you have 3 loops in your form you will have therefore 3 additional sheets. To analyse such data you can use Power BI or Excel vlookup() functions (or the combination index(match())). To have a better idea on how Repeated data is displayed in a database and how to import data from KoBoToolBox to Power BI you can see this video tutorial : Import “Repeat” Data to Power BI - using data export from KoBoToolbox.
  • If you have GPS coordinates/geographical data: GPS coordinates (or more widely data that you can represent on a map) can be used in both offline and online tools, so long as the tool has a mapping component. Most MDC tools today directly have a map feature to visualize the data, that can be more or less advanced (you might be able to view the data in different colours depending on a criteria), but a lot of the other online tools you can connect to through an API or by importing data will also have a map feature (Power BI, Gogocarto, uMap, ArcGIS online, Google Fusion, Python, etc). Offline tools much used for the mapping component include Google Earth (for very basic visualization) or mapping software such as QGIS or ArcGIS.
  • Multiple languages: If some of your data (free text questions) come in multiple languages, you can use some Google tools to automate their translation online (for non sensitive data of course!). You can see this tutorial using both Google Sheets and Translate.

What is the type of technical/technological environment you are working in ?

The technical environment in which you are will most probably also have a big impact on your choice of tools. The most obvious point will be in terms of connectivity (if you have low connectivity, you will most probably not be able to use online websites to do your analysis). You may also have organizational constraints: make sure that the tool is inline with any guidelines or IT organization constraints that might be existing (i.e. licences available, developing an Excel with macros when macros are automatically deactivated on the laptops used by your organisation…). Beyond that, you may also have data protection reasons (in relation to your policy, the legislation you need to follow…) that will orient you towards one tool more than another based on the sensitivity of the data you are analyzing and sharing, or the where the data is stored.

What are the skills and time available for the teams analysing the data?

Make sure the tool(s) you choose:

  • can be set up and maintained by your teams: i.e. an advanced online dashboard requiring more advanced skills is not a good idea if you expect a heavy turnover in your teams and therefore you cannot ensure that your tool will be sustainable in the long run. If you need (or want) an advanced tool for which you are not sure you have the required skills, discuss with HQ to find an external solution.
  • can be used by your teams: the final user of the tool (program and project managers, M&E teams…) needs to feel sufficiently at ease with the tool chosen – for example if the tool set up is on R or SPSS, tools on which only a few staff members are trained, then it might be tricky to ensure that the final users are comfortable enough to do their analysis

What is the budget ?

Many analysis tools have licencing costs or may require a dedicated budget (through externalization for example) to set it up. Is this budget available or do you wish to keep things internal? Make sure you consider this as early as possible in the process.