FinGraph is a free, single-page web application that unifies exploratory network analysis and a broad suite of econometric and accounting-research methods for firm-year financial-statement data, running entirely in the user's browser so that data never leave the machine. It auto-recognises major provider conventions (Compustat/WRDS, CRSP, CSMAR, Refinitiv Worldscope/Datastream), derives ratios and distress scores, and supports correlation and partial-correlation (Gaussian graphical model) networks with FDR/backbone filtering and community detection; firm peer networks; panel regression (pooled OLS, fixed and random effects, Fama-MacBeth) with cluster-robust standard errors; a diagnostics suite and a robust fixed-vs-random-effects (Mundlak) test; moderation and mediation; quasi-experimental designs (difference-in-differences/event study, IV/2SLS, propensity-score matching, regression discontinuity, Callaway-Sant'Anna staggered DiD); Altman/Ohlson/Piotroski/Beneish scores; Jones and modified-Jones discretionary accruals and a Benford's-law screen; principal component analysis and k-means firm clustering; quantile regression and specification-curve robustness; and machine-learning prediction. Every estimator was validated against data generated from known processes and against standard reference formulae. Available free at https://draistudio.com/fingraph/ under the MIT License.
Aneeq Inam (Sun,) studied this question.