A flat map makes certain relationships look obvious. Two road lines meet at the same place on the page, so it is tempting to treat that place as a junction in a planning network. Yet one road may pass above the other without a connection. The drawing can be accurate as a picture of their horizontal positions while the inferred network is wrong. This is not merely a technical detail for a specialist to repair after release. It changes the movements that an analytical product represents as possible.
Our fictional 2022 review examined eighty-six selected crossings. Fourteen were grade-separated without a connection between the roads, while seventy-two were actual junctions in the reviewed evidence. A proposed account treated all the visible crossings as connected. The resulting map could look tidy and complete. Its problem was not obvious from the visual finish, which is why the network relationship needed a review of its own. I recommended holding the affected account and committed ₹53 lakh to the work with survey colleagues rather than rely on a broad qualification attached to the release.
The drawing can conceal the relationship
Spatial products answer several kinds of question. A line can describe where a road lies in a representation. A network relationship describes how a user or an analysis may move from one element to another within the model. A restriction can further limit a movement that a physical connection otherwise permits. Those meanings are related, but none should be inferred automatically from the others. A visually plausible intersection does not establish connection, and a connection does not establish that every turn or every vehicle use is permitted.
The distinction becomes commercially important when a product is sold for planning rather than simple reference. The buyer may use its network to compare options or identify a route for further investigation. An invented connection can affect that analysis even if the product never claims to provide final operating instructions. The enterprise still owes an accurate account of the relationships within the scope it offers. Saying that customers must exercise judgement does not remove the supplier's responsibility to correct a known modelling error in the selected data.
The opposing concern deserves a fair account. Review takes time, and customers may be experienced enough to spot some obvious inconsistencies. A delayed release also has a cost. But experience should not be treated as a substitute for evidence that the supplier already knows is needed. A general caution can explain a genuine boundary of use; it cannot make an unconnected crossing into a connection. The right question is whether the product can be released with a narrower, honest scope or whether the affected relationship must be resolved first.
Correct the selected network without claiming the whole journey
In our review, removing every crossing would have avoided the fourteen false connections while also destroying seventy-two genuine junctions. That would have traded one distortion for another. The work retained the actual distinction and the source basis for it. The model needed to represent which relationships were supported, not merely become less connected everywhere. A data-quality response should preserve useful evidence as well as remove an unsupported inference. Otherwise the apparent correction can make the product simpler while making it less faithful to the setting it describes.
The selected population also mattered. Eighty-six crossings were reviewed; the result did not establish that every relationship in a larger network was correct. A complete-looking product should not borrow the verification status of its reviewed subset. Nor did the seventy-two junctions establish all turn permissions, current conditions or vehicle suitability. Those questions require their own evidence and responsible owners. Our outcome was a corrected relationship account for the defined review, with the remaining scope explicit. It was not a certification of a journey or an instruction to use a particular road.
That boundary should appear in the commercial and executive account, not only in a technical appendix. The ₹53 lakh programme funded relationship review. It did not by itself save fuel, shorten a route or improve safety, because those outcomes were not measured in the case. A more faithful network may enable better subsequent decisions, but an enabling condition is not the same as a demonstrated operating result. Keeping that distinction prevents the cost of a necessary data repair from being justified by benefits the enterprise has not established.
Give the product a responsibility it can actually carry
A useful product brief should identify the decisions the data are intended to support. If the offer is a reference picture, it needs an appropriate account of the features and their sources. If it is a network for planning, it also needs relationships suitable for that analysis. If a customer wants actual movement approval, the conversation reaches responsibilities and evidence that may sit outside the supplier's role. Defining those differences early makes it possible to price and fund the right work, rather than discover at delivery that a familiar word such as route has been used for several different promises.
The data leader's role is to connect those promises to the model and its evidence. I would ask survey colleagues what was observed, product colleagues what the user will infer and commercial owners what has been offered. I would not ask one function to settle every issue alone. A qualified survey finding may resolve a connection question while leaving a current restriction or actual access decision elsewhere. The enterprise needs those responsibilities to fit together without allowing a convenient data field to become an unofficial approval mechanism.
For directors, the case raises an investment and accountability question. What capability is management funding, and what claim will the enterprise be able to make when the programme is complete? A board should not need to inspect line geometry or choose a software setting. It should understand whether the release depends on relationships that have been examined, whether unreviewed scope remains visible and whether the commercial description matches the result. That scrutiny can be precise without turning directors into network engineers or bypassing the operating owners of the actual work.
The broader lesson is to distrust visual completeness as a substitute for semantic completeness. A polished map can encourage users to supply relationships the evidence has not established, especially when the product's presentation makes every feature appear equally settled. Our fourteen grade-separated crossings were a small selected group, but they exposed a larger question about what the enterprise believed it was selling. I want the model to carry a responsibility it can support: a defined account of spatial relationships, with the evidence and remaining decisions visible, rather than a picture whose apparent connections promise more than the road provides.