davvvNET Solutions — analytics, integrations, and custom engineering
Request a demo Schedule a meeting

Data / Finance 150

Finance 150 · Institutional market analytics

Market evidence, built to withstand scrutiny.

A governed market-data architecture and liquidity-focused Top 150 framework turn unusual activity into explainable analytical outputs—with the identity, completeness, lineage, and calculation controls investment teams need for defensible review.

Discuss the analytics capability Review the evidence path

01 / Trust architecture

Evidence before inference.

The architecture preserves what arrived, proves what was complete, and records which version produced each result. That discipline makes analytical revisions controlled and reviewable instead of silent.

  1. 01 Source evidence Immutable source records, receipt verification, collection context, and delivery completeness establish what was received. An attributable source record for audit and reconstruction.
  2. 02 Normalization Stable security and listing identities, dated ticker assignments, and exchange-session calendars handle symbol reuse cleanly. Comparable observations without treating a ticker as permanent identity.
  3. 03 Quality proof Gap detection, controlled recovery, coverage checks, and documented completeness establish readiness before calculation. Fewer silent data-quality surprises and explicit reasons for withheld results.
  4. 04 Analytics Prior-only baselines, effective-time and knowledge-time controls, and reproducible calculation versions prevent quiet look-ahead. Defensible comparisons under a declared methodology.
  5. 05 Governed output Input lineage connects each analytical result to its source evidence; incomplete coverage fails closed. A reviewable output—not an unexplained score.

02 / Focused universe

Why Top 150

A prospectively selected set of 150 investable U.S.-listed securities concentrates analytical review on names with meaningful, sustained activity. The focus improves relevance and reviewability; it does not force 150 results. A security appears only when observed measures cross transparent thresholds and every required integrity check is satisfied.

  • Trailing median dollar volumeRanks typical activity without letting one exceptional session dominate.
  • Consistent session coverageMissing or invalid sessions do not become zero-volume observations.
  • Price and investability checksEligibility rules are fixed before the constituent period they govern.
  • Lower-tail resilienceLiquidity review considers weaker activity periods, not only the median.
  • Controlled turnoverProspective effective dates keep constituent changes explicit and reproducible.
  • Benchmark contextDeclared benchmark instruments remain outside the 150-name quota.

The broader research universe and historical constituent versions are retained so prior comparisons can be reproduced, reviewed, or revisited without rewriting membership from today’s roster.

03 / Analytics framework

Unusual activity, plainly defined.

Each measure answers a specific review question. Qualification can come from one or more visible components; there is no required opaque composite. Unusual activity is an investigation signal, not a forecast of future return.

  1. 01

    Volume Surge

    Current-session volume divided by the median of valid prior sessions. It highlights participation that is elevated relative to the security’s own recent baseline.

    Qualifies at 1.5× or more

  2. 02

    Trading Activity Surge

    Current transaction count compared with its prior-only median. It separates more frequent trading activity from volume alone.

    Qualifies at 1.5× or more

  3. 03

    Unusual Price Range

    Normalized true range is compared with its recent median and median absolute deviation, capturing an atypical high-low range or gap-adjusted move.

    Qualifies at |robust score| ≥ 2.0

  4. 04

    Price Change Context

    Close-to-close movement is shown beside activity measures to aid interpretation. By itself, price movement is not treated as a recommendation.

    Current close ÷ prior close − 1

  5. 05

    Relative & benchmark context

    Declared benchmark-relative return and broader-universe context can distinguish security-specific movement from a shared market backdrop when their inputs are qualified.

    Security return − benchmark return

04 / From observation to evidence

An explainable result carries its own proof.

This conceptual record shows the composition of an analytical result without implying a market observation or performance outcome.

  1. ObservationOne stable security and listing identity at one declared exchange session.
  2. ComponentsVolume, transaction count, normalized range, and price-change context are evaluated separately.
  3. ThresholdsEvery crossed measure is named, with its comparison basis and distance from threshold.
  4. IntegrityCoverage, valid prior history, knowledge cutoff, benchmark availability, and input lineage must pass.
  5. DispositionQualify for investigation when the rule and gates pass; otherwise withhold with an explicit reason.

05 / Integrity & reproducibility

Built for the second question.

Due diligence rarely stops at “what did the measure show?” The architecture is designed to answer what data existed, which rules applied, what changed, and why a result was included or withheld.

Point-in-time identity

Security, listing, ticker assignment, and constituent membership are dated for when each was valid, so symbol reuse and later roster changes do not rewrite prior identity.

Controlled recovery

Repeatable processing and targeted gap recovery restore coverage without duplicating accepted facts or silently closing unproven gaps.

Information cutoff control

Inputs must have been available by the declared knowledge cutoff; later corrections create a new analytical version.

Completeness status

Observed, continuous, and verified-complete status distinguishes “latest seen” from evidence that a defined period is complete.

Calculation versions

Formula, lookback, threshold, benchmark, universe, and missing-data changes create a new methodology version.

Input lineage

An analytical result retains the source edition, evaluation time, universe, completeness status, and methodology version used to produce it.

Quality-first exclusions

Incomplete required coverage, invalid denominators, insufficient history, or missing context produce explicit exclusions—not inferred values.

Reproducible evaluation

Constituent versions, cutoffs, calculation versions, and retained research history support defensible historical comparison.

Publication principle

Coverage and quality evidence are part of the analytical contract. When completeness is not proven, the affected analytics are withheld.