Difference between revisions of "Data Analysis"
Kbjarkefur (talk | contribs) |
|||
| Line 4: | Line 4: | ||
== Read First == | == Read First == | ||
Data analysis typically has three stages: | |||
#Exploratory Analysis | |||
#Final Analysis | |||
In exploratory analysis, emphasis will be on producing easily understood summaries of the trends in the data so that the reports, publications, presentations, and summaries that need to be produced can begin to be outlined. Once those stories begin to come together, the code is re-written in a "final" form which would be appropriate for public release with the results. | |||
== Preparing the Data Set for Analysis == | == Preparing the Data Set for Analysis == | ||
See [[Data Cleaning]]. | |||
In the cleaning section we do replace values that otherwise would bias the variable. The first part of Data Analysis is to edit the variables so that they fit into the statistical analysis models that we are using. | In the cleaning section we do replace values that otherwise would bias the variable. The first part of Data Analysis is to edit the variables so that they fit into the statistical analysis models that we are using. | ||
Revision as of 19:36, 6 November 2017
Read First
Data analysis typically has three stages:
- Exploratory Analysis
- Final Analysis
In exploratory analysis, emphasis will be on producing easily understood summaries of the trends in the data so that the reports, publications, presentations, and summaries that need to be produced can begin to be outlined. Once those stories begin to come together, the code is re-written in a "final" form which would be appropriate for public release with the results.
Preparing the Data Set for Analysis
See Data Cleaning.
In the cleaning section we do replace values that otherwise would bias the variable. The first part of Data Analysis is to edit the variables so that they fit into the statistical analysis models that we are using.
- Standardization - Convert all values in each variable into the same unit. If the values in one variable are different then errors like 1000 gram will be interpreted as one thousand times larger than 1 kg.
- Aggregation - We often collect variable disaggregated over categories (income collected as different income categories) or disaggregated over instances (harvest value over multiple crops). Disaggregated data collection is used to improve quality of data collected, but in the analysis we are often interested in the aggregated value.
Outputting the Result of the Analysis
Just as the rest of your code the output of results must also be replicable. There are different degrees of replicability. The basic that is obviously a must is that all parts of the results used in the table is replicable.
Even better is that all part of the same table is outputted in a single file. Sometimes tables are consist of results from multiple estimations and it is preferably that they are outputted to a single file. See Stata command estout.
Optimally all tables are outputted in a way that no manual formatting is required. A very common tool for that is LaTeX. DIME has prepared material for getting started with LaTeX that assumes no knowledge in LaTeX and aims to explain the work flow from software as Stata and R to final reports using LaTeX. [[1]]
Different Specific Types of Analysis
Principal Component Analysis
Principal Component Analysis (PCA) is an analytical tool looks to explain the maximum amount of variance with the fewest number of principal components.
Cost Effectiveness Analysis
One type is Cost-effectiveness Analysis