SenSight is a complete statistical quality engineering toolkit that runs on your phone, tablet or desktop — and keeps working when the network does not.

For engineers who decide on the shop floor: ship this batch or not, is the process in control, can this gauge be trusted.

PROCESS CONTROL (SPC)
• I-MR, X-bar R, X-bar S, EWMA and CUSUM charts; p, np, c and u attribute charts
• Subgroups by fixed size or by a grouping column (shift, batch, machine)
• All eight Nelson rules; unequal subgroups reported, not averaged away

PROCESS CAPABILITY
• Cp/Cpk, Pp/Ppk and Cpm with the Anderson-Darling normality test built in
• Non-normal data: ISO 22514-2 percentiles and Box-Cox — the panel states which number to report

MEASUREMENT SYSTEM ANALYSIS (Gage R&R)
• Crossed Gage R&R by AIAG average-and-range and by ANOVA; nested designs
• Bias, linearity and attribute agreement, with a warning when the design cannot answer it

DESIGN OF EXPERIMENTS
• Factorial and screening designs with interaction plots
• Response surface (central composite, Box-Behnken) — saddle points are named, not sold as optima
• Taguchi L4-L27 with S/N ratios and confounding stated; mixture designs

SIX SIGMA QUALITY TOOLS
• FMEA with AIAG-VDA action priority, not just an RPN number
• Control plans with reaction plans, carried over from the FMEA
• Root cause: fishbone and 5 Whys, checked for stopping at human error; project charter

REGRESSION & MODELLING
• OLS with transforms and full diagnostics; best-subsets and stepwise selection
• Poisson with overdispersion detection; curve fitting across eight shapes; logistic regression
• Tree-based models: CART and random forests for thresholds and interactions a line cannot see
• What-if sliders: move an input, watch the prediction move

MORE ANALYSES
• ANOVA with Tukey, t-tests, chi-square, non-parametric tests
• PCA, k-means with silhouette scores, discriminant analysis
• Time series: Holt-Winters and ARIMA, measured against a naive benchmark
• Reliability with right-censoring: Weibull/lognormal maximum likelihood, Kaplan-Meier, B10/B50 life
• Acceptance sampling with OC curves, tolerance and equivalence tests; Likert, NPS, Pareto and 16+ charts

ONE REPORT FOR EVERYTHING
• Every analysis recorded automatically and saved on your device
• Export as one Excel workbook (a sheet per analysis, charts embedded) or printable HTML
• Your own branding; hash-chained audit trail for regulated environments

FIELD & FACTORY DATA
• Offline: loading data, the main regression and the Six Sigma tools need no network; server statistics resume when you reconnect
• CSV/Excel import, OCR from a photographed table, or direct reads from PostgreSQL, MySQL, SQL Server, Oracle, SQLite
• Live data from Modbus TCP/RTU, OPC-UA, EtherNet/IP, Siemens S7, BACnet/IP, DNP3, MQTT and BLE — and write coefficients back to the device
• Historian with roll-ups that keep the peaks; REST API and an offline-queue field app
• Decimal comma support for Turkish and European formats

BUILT TO BE HONEST
It says so when the data is not normal, when a stationary point is a saddle rather than an optimum, when a forecast cannot beat the naive benchmark, and when a Gage study cannot answer the question. 47 lessons explain the method behind each result. 12 languages, right-to-left included.

PRIVACY
Analysis runs on our server: the numbers you analyse are sent there to be computed and are not kept — a dataset is stored only if you save or share it. The main regression runs on your device. With an account we store your profile, subscription and saved sessions. Detail: https://sensighttech.com/privacy

PRICING
Free: 500 rows, regression and charts. Standard adds statistics, time series and multivariate. Pro adds SPC, MSA, DOE, Six Sigma tools, the historian, database connectors and two live industrial connections. Enterprise: unlimited connections, SSO, multi-user. Subscriptions renew automatically; cancel any time from your account.

Support: info@dropless.com.au