The Vidya Samiksha Kendra, housed in Pune under Maharashtra’s School Education Department, was established with a genuinely ambitious purpose: to aggregate fragmented data streams from across the state’s vast school system into a unified analytical hub and to make that intelligence available to administrators from the district level downward. On paper and very much in the promotional literature surrounding its launch, it is a serious attempt to move Maharashtra from compliance-oriented school monitoring toward something more honest and more useful — governance driven by evidence.
A recent visit to the VSK in Pune, followed by school-level field observations in Nashik district, produced a more complicated picture. The infrastructure is real. The intentions appear genuine. But the gap between the system’s design and its actual functioning on the ground is wide enough to raise a question that those invested in its success would do well to sit with: can a data centre drive educational governance when much of the data it needs is absent and when those who are meant to use it don’t quite know how?
What the VSK has and what it doesn’t
The VSK currently draws on two primary data sources: Periodic Assessment Test(PAT) results and daily school attendance. Both feed into a central dashboard that, in theory, allows officials to monitor student learning outcomes and physical presence in near real time. This is not nothing. Maharashtra’s school system encompasses approximately 2 lakh schools and more than 2 crore students spread across 36 districts and even partial visibility into such a system at scale is a non-trivial achievement.
But the system is conspicuously missing data from several frameworks that are central to school education governance in the state. UDISE+ records, the Performance Grading Index(PGI), the SQAAF quality assurance framework, NIPUN Maharashtra learning benchmarks and basic entitlement data on textbook and uniform distribution — none of these are integrated into the VSK dashboard. For a platform that describes itself as a unified education intelligence system, these are not minor omissions. They are the difference between a partial instrument and a comprehensive one.
The gap between aspiration and data availability was apparent in conversations with district and block-level officials during the Nashik field visit. These officials — District Education Officers, Block Education Officers, Kendra pramukhs and Extension Officers — reported that while PAT data is technically available on the dashboard, district-level analyses of student performance are not accessible to them directly. To obtain any disaggregated picture of how students in their jurisdiction are performing, they must write a specific request by email and wait. This, in a system designed to enable real-time, data-based decision-making, is a significant contradiction.
The data literacy gap: Officials who cannot read the room
Even where data is available, there is a deeper problem. A majority of the officials who are expected to use the VSK — Kendra pramukhs, Extension Officers and BEOs at the block level— lack meaningful data literacy. They know the dashboard exists. Several have received brief orientations. But extracting data, interpreting trends, forming questions from what they see and then making administrative decisions on that basis remains largely beyond their current practice. The consequence is that officials often depend on a third person, typically a technically inclined subordinate or an informal contact, to pull numbers from the system for them. This intermediation is not just inefficient; it defeats the purpose of decentralized access entirely.
There is no shortage of evidence in the literature about how this plays out. Data systems that are technically sophisticated but operationally inaccessible to their intended users tend to produce what might be called performative compliance — officials acknowledge the system, generate the expected outputs and then continue making decisions the way they always have. What is needed and what appears to be largely absent in the VSK’s current implementation, is a sustained effort to build data fluency among district and block functionaries. YouTube live sessions and Zoom walkthroughs, offered at the right intervals and in Marathi, could accomplish a great deal more than occasional in-person workshops that few attend and fewer remember.
The attendance problem: 98 per cent present, reality unknown
Perhaps the most striking observation from the field concerns attendance data — the one indicator that the VSK tracks every single day. Maharashtra’s school attendance dashboard routinely shows figures in the range of 95 to 98 per cent. This would mean that nearly every enrolled child is in school, every day. Anyone who has visited government schools with any regularity knows this does not reflect reality.
The explanation is structural. School attendance in Maharashtra is also the basis for Mid-Day Meal headcounts, which in turn determines the quantity of food cooked and the amounts disbursed. The incentive to mark full or near-full attendance is therefore not an individual teacher’s dishonesty but a systemic logic — one that the VSK’s current design does nothing to disrupt. A teacher can mark 100 per cent attendance on the app without any verification. No server-side check flags improbable figures. No algorithm compares today’s entry against the historical average or against the physical capacity of the classroom. The data flows in, and the dashboard accepts it.
This makes the attendance layer of the dashboard not just unreliable but actively misleading. When senior officials at the state or district level look at attendance numbers as an indicator of school functioning, they are looking at a reflection of the Mid-Day Meal system’s reporting logic, not at children’s actual presence in classrooms. Decisions made on this data — about staffing, about intervention priorities and about whether a school needs attention — will be distorted accordingly. The solution being piloted in some quarters, facial recognition-based attendance marking, is worth watching. Its success will depend on consistent follow-through from district-level leadership to ensure the technology is used as intended and not simply routed around.
Data Mistrust: When teachers disagree with the dashboard
A recurring theme in school-level conversations was that teachers do not always trust what the dashboard shows. In several instances, figures displayed on the VSK — related to their students’ PAT scores or attendance records — were disputed by the teachers themselves, who believed the data to be inaccurate or incorrectly attributed. This is not a trivial concern. A governance system built on data that the people closest to that data do not trust is built on sand.
Data validation — the process of checking, cleaning and reconciling data before it is used for decisions — is not a glamorous function, but it is foundational. If a teacher believes her students’ scores have been entered incorrectly, or that attendance figures for her school don’t match the register, and there is no mechanism by which she can flag this, query it, or have it corrected, the system’s credibility erodes. This is the kind of erosion that is slow at first and then sudden — and very difficult to reverse.
Towards a VSK that works: What needs to happen
The VSK is not a failure. It is, as yet, an incomplete project — and one that could become considerably more useful if its structural weaknesses are addressed with the same seriousness that was brought to its technical design. Several things seem necessary.
First, analyzed data — not just raw figures — must be made accessible to district officials directly, without requiring them to write a mail and wait. If the system can generate a report for the state, it can generate one for a DEO. The design choice to restrict this access makes the decentralization argument hollow.
Second, data and dashboard literacy must be treated as a governance priority, not an afterthought. Regular online training — not a one-time workshop — is needed for BEOs, Extension Officers and Kendra pramukhs. The training should focus on how to extract, read, and act on data, with follow-up accountability from higher authorities.
Third, the VSK must integrate the data it is missing — UDISE+, NIPUN Maharashtra, PGI, SQAAF, entitlement distribution records — or it will remain an instrument that sees only a narrow slice of the school system it is meant to serve.
Fourth, the government needs to invest in permanent human capacity at the VSK itself. Data analysts, education specialists and domain experts — on regular posts, not outsourced arrangements — are what will give the centre the ability to interpret and communicate what its data is saying. A dashboard without analysts is a car without a driver.
Finally and perhaps most importantly, the state government needs to articulate a clear and sustained vision for data-based education governance — one that comes with timelines, accountability structures and visible leadership from the top. Data-based decision making does not take root because a platform exists. It takes root because those in authority consistently ask data-based questions, demand data-based answers, and act on what they receive.
There is a phrase that came to mind repeatedly during this visit: a toothless tiger in the making. The VSK has the form of a powerful accountability instrument — the servers, dashboards, the institutional mandate. What it has not yet developed is the bite. The data is partial. The users are underprepared. The feedback loops are broken. The attendance numbers are almost certainly fictional.
None of this means the project should be abandoned. It means it needs to be taken more seriously than it currently is — by the government that funded it, by the officials who were given access to it and by the education system that was supposed to be transformed by it. The VSK is the right kind of institution for the problem Maharashtra’s schools face. Whether it becomes genuinely useful will depend on decisions made in the next few years, not about technology, but about commitment.
