PT Data Migration
Migrating legacy data
Overview
This section covers the principles eGov will be following for migrating legacy property registry along with demand and collection balances to the DIGIT system (currently in Uttarakhand but can be used for other clients in future). eGov has been successful in implementing property tax applications in states of Andhra Pradesh, Punjab and various other Municipal Corporations like Greater Chennai, Nagpur and so on using best practices defined here.
Entry Criteria
The state team is supposed to take care of the below activities before property data is shared to eGov for data migration:
State team to share all master data and its localization that are part of the property tax. For example masters like - city, boundary, usage type, property type and so on
State team to do the data gathering and prepare the property tax data in the predefined template shared by eGov
Data prepared needs to be scrutinised by the data owner from the State’s team. Following data validations to be considered in particular:
Validate if all sheets have accurate records (owner details, property details and DCB details, Floor details) in terms of what is expected
DCB details can be more than the number of records in the property sheet
There can be multiple owners for a single property. But there needs to be an owner and DCB for each of the properties
Have a check on all mandatory data present
All sheets to be aligned with the unique id field
The state has to share the complete set of property data for a city in a single go. They can also give one set for residential properties and another for commercial properties, but cannot give multiple sets for the same.
Data owner to give a sign off on data before handing over to eGov team
State team to be aware of assumptions and default values that will be applied as part of the data migration process. This will be shared by the eGov team
UAT and production environment where the data is to be loaded to be provisioned and set up
Principles
Step 1 - Data cleansing
Data received from the State team will be cleansed by eGov with default values
Mandatory fields that have no financial impact on citizens can be filled with default values in the absence of values specified by the State team. Few examples are listed below -
Covered areas can be updated as 0
Road type as 12 meters
The number of floors can be updated as 1
Certain fields that have a financial impact can also have default values like tax amount in the absence of values specified by the State team. (It can be updated as Rs.1)
If the mobile number is blank, a dummy mobile number will be updated. Care will be taken to ensure that this is not an actual mobile number and this number will be removed from notifications. Dummy numbers can be something like 99999xxxxx where the first few dummy numbers in a state will be 9999900001, 9999900002, 9999900003 so on.
Follow the below assumptions to do data cleaning
If the mobile number and user name combinations are duplicating we will consider these properties belong to the same owner. This means one user will be created in the system which will have a link to two properties.
Have a marker if any assumptions and default values are applied on a record
Certain data are currently not assumed or defaulted. Below listed are a few cases -
Owner details completely not available for a property - If complete owner details are missing, these records to be kept in the not to be migrated list.
If DCB is not there or given for a range of years - since there are a lot of use cases here, we can keep this record in the not to be migrated list
Duplicate old property id - leave the data as it is and then edit it later using the data entry screen
If boundary data is not present or data not aligned with the master data shared - update to a default mohalla which the state agrees to
Door number is not present - update the value as “door number”
Owner Name blank in the owner details sheet - update the value as “owner”.
Step 2 - Data Load on UAT
Load the records using property migration kit to UAT server
Prepare a list of all passed and failed property records and share them back with the State team. Passed records along with values defaulted will be shared.
If the State team wants eGov to retry all failed records they are expected to come back with the revised data within one week.
Repeat Step 1 for revised records received from the State team.
The final status of the data load will be shared with the State team.
Step 3 - Move data to Production
State team to verify data loaded on UAT server and provide a sign off for each of the ULBs within 1 week
On getting sign-off of UAT data, eGov will load that data to production using the migration kit used for loading to UAT instances. This can be done only if the production infrastructure is ready.
Get sign off on the production data before the application can be made live.
References
ULB | Total Records | Duplicate/Wrong data | % of data loadable after imposing assumptions | % of incorrect data |
Roorkee | 16363 | 500 | 97% | 3% |
Rishikesh | 2877 | 356 in the first sheet 261 in the second pass | 88% 91% | 12% 9% |
Communication Mechanism
Communication for data migration exercise can be split into the following stages:
There will be a SPOC from the State team who is the data owner who will share data with the eGov team. Excel sheet data will be put on google drive and shared with the project manager from eGov. Similarly, all communications from the eGov side to State will be routed through the project manager.
Pre-migration
A demo of how PT module works - State + ULBs
Product fitment discussion, configurable data finalization and sign off - State
An overview of how data migration works - State + ULBs
Sharing the template (along with checklists and FAQs) in which we will collect data and type of data required against each field along with training - State + ULBs
Sharing assumptions which will be followed while migrating data - State + ULBs
During migration
Data inconsistencies observed in Data cleaning stage - State
Gaps needed to complete data migration (like mohalla mapping) - State
List of records which will be migrated as-is and the ones which will be migrated with assumptions - State
Post-migration
List of properties migrated without error - State
List of properties migrated with error for ULBs manual entry - State
UAT Sign-off document - State + ULB
Final confirmation post migration on Production - State + ULB
Way Forward
All failed records from the second pass will be manually entered into the system using data entry screens
The state has to make a mandate to collect property tax only using DIGIT.
All records with assumption markers or incorrect data can be updated by counter employees when citizens come for payment. Maybe this option can be given to the citizen when coming for online payment.
When a citizen is coming for making his payment, the counter operator can check if their property exists in the system using propertyid or mobile number. If it is found to be missing, it can be captured afresh in the system using the data entry option. All required information can be obtained from the citizen, either from previous year receipts that they carry or directly asking them.
With this approach, a ULB can go live with 90% of property data within a short time.
Data Mapping Template
Data Migration Checklist
Data provided in the standard template attached only will be accepted for further analysis and process.
Data provided into the template is loaded into staging tables and then validation and data clean-up is done.
Properties with current Tax 0. (2020-21)
Properties with advance collection
Properties with current tax partially payment
Properties with Mohalla not configured in master
Masters mapping with the configured masters data in the system.
All the findings of data are reported to the onsite team and sought clarification.
Once all the queries/ findings are cleared/addressed data is moved into UAT.
Data is verified into UAT with sampling by the state team. Go ahead.
Preparing some basic checkpoints before moving into production.
Move the data into production and verify the checkpoints.
Last updated
