Data Processing
We perform a Data Cleaning process in 3 steps:
1-DATA BASE INCONSISTENCIES DETECTION
We conduct a thorough analysis of consistency for all variables// of variables consistency in the dataset. All the checks are performed regarding the constraints defined in the questionnaire. The principal checks are:
- Single variable consistency checks.
- Cross variable validation checks (multiple variables).
- Skip question checks.
Depending on the particular characteristics of each project, we may perform additional controls or avoid some of those mentioned above.
2 – CORRECT VALUE SEARCHING IN THE DATA COLLECTION INSTRUMENTS
After detecting all the inconsistencies in the database, we need to find the correct values in the original instrument used in during the data collection.
If the Data Collection Process has been done by a third party, then the entity who has performed it must respond to us with the required answers.
3 – CORRECTION ON THE DATA BASE
After receiving the document with the responses, we will fill in missing values, leave or remove outliers, and solve inconsistencies on the dataset. We delivery a log of all the actions (changes) performed to the dataset.
We can merge with the working dataset any external data you may have. Datasets with different levels of analysis units can be merged as individual information with household information, etc.
You can also be interested in merging information from biological or chemical analysis results from a Lab, or any other information coming from a different source.
As part of the preparation process data, at this stage all necessary indicators which will then be used in subsequent statistical analyzes are built. We calculate standard indicators as well as any other indicators where the customer provides the necessary information and instructions.
See some examples for Nutrition area:
- «Indicators for assessing infant and young child feeding practices» (IYCF) from WHO and UNICEF last series recommendations.
- «Household Hunger Scale» from the Food and Nutrition Technical Assistance III Project (FANTA)
- Anthropometric indicators. – length / height-for-age, weight-for-age, weight-for-length, weight-for-height and body mass index-for-age – from WHO Child Growth Standards.
- Wealth Index recommended by the World Bank and WFP.