Here are some points to consider when planning a data analysis project:
| • | Decide on your objectives and identify the key data elements you will need. Identify your lines of inquiry or prepare a list of questions you want to answer. |
| • | In consultation with the data owner and systems specialist plan the best way to acquire the data. Identify which data files, records and fields you need. What data storage format will it be in? Will it need to be on tape, disk or accessed through a computer network? If the files are large, you may want to have them sorted first. Consider using ODBC for more efficient data access. |
| • | Make a formal written request for the data. Request a record layout for each record type: this may include specific information identifying different record types and their relationships. Request the data in your preferred format: disk, tape or ODBC access. Request control information such as record counts and hash totals. It may be useful to work with a small sample file to test your procedures before requesting the complete file. |
| • | Obtain the data together with record layouts and control totals, if necessary. |
| • | Use the Data Definition Wizard to create table layouts so Analyzer can read and display the table data. |
| • | Verify the integrity of the table data. Use tools and the record layout in Analyzer to ensure you have the correct files, the correct number of records and that the control totals agree. Verify that all data is valid. Look for inconsistencies between the data you have and your knowledge of the business area. Discuss unexpected results with the data owner. |
| • | Use commands and functions to explore the data and answer your questions. Use the Command Log and View notes to document your work as you proceed. |
| • | Confirm your findings with those who can verify them. Have your manager or co-worker review your procedures. Explore possible reasons for anomalies that you discover. Make sure that your assumptions, conclusions or recommendations are valid. |
| • | Discuss your findings with the data owner and other interested parties and perform further analysis as required. |
| • | Document your findings in a report. |
| • | Archive your Analyzer Project and related data for future reference. |
| • | If there is a need for regular, on-going analysis, consider automating your analysis using Analyzer’s procedure capabilities. |