00:00:04 Introduction and series recap

Hello everyone and welcome again. This is the fourth session in our webinar series. In our earlier sessions, we embarked on a journey to transform our MEAL plan into actionable data insights for the Food for Progress project. In session one, we transformed the MEAL plan into a comprehensive data model, identifying key indicators and data requirements. In session two, we built a database in ActivityInfo based on our data model, structuring forms and fields to capture the necessary data. In session three, we focused on managing and processing data within the system, ensuring data quality through validation rules and consistent data entry practices. Today, we will focus on analyzing the data we've collected and visualizing our findings to support decision-making.

I am Firas, Implementation Specialist at ActivityInfo, and I want to thank my colleague, Eliza, our Senior Implementation Specialist, for co-hosting this webinar with me. Here is what we will cover today: First, "From data to insights: a digital journey," where we explore how raw data is transformed into actionable insights. Then we will dive into data analysis, examining the processes and techniques involved in analyzing our data effectively. We will explore data visualization to discuss how to present our findings in a clear and impactful way. Finally, we will have time at the end for questions.

00:01:53 Project context and data model

Let us revisit the context of our project to ground our analysis. The Food for Progress program targets developing countries and emerging democracies to modernize and strengthen the agricultural sector. Our specific project focuses on improving agricultural productivity and expanding the trade of agricultural products in Mauritania. Our MEAL plan is centered around two key indicators. The first is the yield of targeted agricultural commodities among program participants with USDA assistance, calculated as yield per hectare. The second indicator is the number of individuals accessing agriculture-related finance as a result of USDA assistance.

Previously, we developed a detailed data model to support our data collection and analysis efforts. Our data model includes entities like beneficiaries, farms, field officers, and reference data such as administrative levels. There is a natural relationship to everything; field officers are connected to beneficiaries, farms are subsets of beneficiaries, and we have specific collection forms for our two indicators. In terms of reporting, we have two main streams: donor reporting, which focuses on indicator progress on an annual basis, and internal monitoring reports, which track metrics like registered participants and sex disaggregation on a monthly basis.

00:06:03 The data analytics process

Before diving into data analysis, it is crucial to ensure we are fully prepared. We need to define clear objectives and ensure data preparedness by confirming our data is accurate, complete, and reliable. The data analytics process consists of four key steps: data collection, data preparation, data analysis, and data visualization. Each step builds upon the previous one.

When collecting data, we must ensure it aligns with project objectives, is reliable, collected promptly, respects participants' rights, and uses appropriate techniques. Data preparation involves cleaning data to correct errors, transforming data to standardize formats, and integrating data from various sources. Proper preparation ensures our analysis is based on reliable and consistent data.

Data analysis involves several key aspects: descriptive analysis to summarize data characteristics; comparative analysis to identify patterns between groups; correlation analysis to explore relationships between variables; the use of tools like pivot tables; and interpretation to draw meaningful conclusions. Finally, data visualization helps communicate findings clearly. This involves defining the purpose, choosing appropriate chart types, tailoring the visuals to the audience, and incorporating interactivity.

00:11:47 Data analysis and visualization concepts

Data analysis enables us to make evidence-based decisions, uncover insights, assess progress, and identify areas for improvement. Pivot tables are powerful tools for this, allowing us to summarize and explore large datasets dynamically. Understanding the components of a pivot table is essential: dimensions (categorical variables like gender), measures (numerical data like total yield), and filters. By manipulating these components, we can customize our analysis to answer specific questions.

Data visualization is essentially storytelling. To tell a compelling story, we need to know our stakeholders. Internal staff might require detailed operational data, while donors might be interested in high-level outcomes. Selecting the right chart type is crucial. Bar charts compare quantities, line charts show trends over time, and pie charts represent proportions. Maps are valuable for visualizing geographic data and spatial patterns, though they must be accurate and not overloaded with information.

When sharing insights, the report layout should be tailored to the audience. For internal staff, a single report focusing on operational details may suffice. For donors, compelling dashboards providing high-level overviews are more appropriate.

00:20:15 Live demonstration in ActivityInfo

Let us jump into ActivityInfo to demonstrate how we transform our reports. We have our database with reference data, beneficiary registration, and surveys related to our two indicators. We created a single report acting as an indicator tracking table. It shows the baseline, target, and calculated progress. For example, we can see the progress to target for the first indicator is 20%, while the second is 78%. We also included a "deviation from baseline" metric to show performance relative to the starting point.

While a table meets requirements, dashboards make data more engaging for donors. We created a dashboard that includes the same indicator tracking table but adds a logo, a title, and a map showing beneficiary distribution. We also added specific pages for detailed analysis. For the first indicator, we have sex disaggregation, farm size, and a line chart tracking trends in cultivated area and production. For the financing indicator, we have charts showing the type of financing and aid. These reports are dynamic and update automatically as data is entered.

For internal reports, we created a notebook to track operational metrics like the number of participants per month. We can easily add text to explain insights. Creating these reports is straightforward using pivot tables; for example, dragging the "month" to the x-axis and "beneficiaries" to the measure instantly generates a bar chart.

00:27:39 Scenario: Field officer performance reporting

Let's consider a scenario where we notice discrepancies and want to evaluate field officer performance. We might ask questions like: How many beneficiaries are registered per officer? How many farm visits have they conducted? In our data model, this data comes from four different tables. A common challenge in the humanitarian field is fragmented systems where linking this data is difficult.

Using an integrated system like ActivityInfo solves this. In our demonstration, we can create a report to track performance even if the field officer isn't directly linked to every single form. Because there is an established relationship between field officers and beneficiaries, and beneficiaries are linked to the surveys, we can pull data from multiple forms into one pivot table. We can drag the field officer's name as a dimension and then select measures from the 'Beneficiaries' form, the 'Financing' form, and the 'Farm visits' form. This flexibility allows us to generate actionable insights and identify areas where additional training might be needed.

00:34:53 Conclusion

To conclude, effective data analysis allows us to make informed decisions and enhance program effectiveness. Clear and compelling visuals make complex data accessible to all stakeholders. Identifying trends allows us to refine strategies, and continuous monitoring keeps the program on track. ActivityInfo serves as an end-to-end solution for M&E data management, integrating collection, organization, and analysis to streamline the process.

00:35:54 Q&A session

What is the difference between data analytics and data analysis? Data analysis is just one part of data analytics. Data analytics covers the whole lifecycle process, from data collection and preparation to analysis and finally visualization.

What is the purpose of adding a column for deviation from baseline? While "progress to target" shows how close we are to the goal, "deviation from baseline" tells us how much we have improved or declined compared to the starting point. It is useful for showing percentage increases or decreases, though it depends on the specific activity and report type.

Do donors usually ask for access to the database? Generally, donors do not ask for access to the primary data or the database itself. They are interested in the summary reports, high-level outcomes, and impact assessments. You can share dynamic reports or dashboards with them without granting access to the underlying confidential data.

Does ActivityInfo help do analysis of correlation? ActivityInfo does not have a specific "correlation analysis" button, but you can achieve this through calculated measures or by observing data. For example, you can place two charts next to each other to observe relationships, or use formulas to calculate specific metrics.

How does the map work? Is it linked to Google Maps? The map functionality is built-in within ActivityInfo; it is not linked to Google Maps. You define coordinates using a Geographic Point field in your form. Once the data is there, you can add a map widget to visualize the distribution of records.

Can we use ActivityInfo for different projects at the same time? Absolutely. You can structure your database to handle multiple projects. In our example, we defined a project form and linked activities and indicators to that project. You can disaggregate data by project, sector, or country, making the system scalable for multiple interventions.

How do you handle qualitative analysis? Qualitative analysis is challenging but possible. You can use the system as a repository for qualitative data (text fields, attachments). To analyze it, you can try to categorize open-ended responses using multiple-choice fields where possible, or use search formulas to find specific keywords within text data. Over time, you might identify patterns that allow you to structure these questions better in future surveys.