Skip to main content
CARLA HQ
Human Resources & Workforce Analytics Operations & Workforce Analytics use casesNACE O84.11RegressionTarget: current_annual_salary_usdOpenML #42125 ↗

Employee Salaries

The Employee Salaries dataset provides compensation records for 9,228 active, permanent public sector employees in Montgomery County, MD. The primary predictive objective is tabular regression to estimate current annual salary based on departmental placement, job titles, tenure, and historical earnings.

✨ Try in CarlaInstall Chrome Extension ↗100% In-Browser WebGPU • Zero Cloud Upload
9,228Records (Rows)
10Predictive Features
4 / 9Numeric / Categorical
9.3%Missing Value Ratio
8,126RMSE
36.28sWebGPU Mean Split
BUSINESS CONTEXT & OBJECTIVEPublic Sector & Human Capital Management (O84.11)

Business Objective: Employee Salaries

01Business Context

Public sector human resources departments and enterprise compensation teams manage complex salary structures across diverse agencies, divisions, and union-negotiated pay bands. Ensuring fair compensation, standardizing career progressions, and budgeting across municipal departments requires accurate salary modeling.

02Analytical Objective

Predict the base current annual salary for municipal employees using demographic, positional, departmental, and historical tenure attributes.

03Economic & Decision Impact

Improves payroll forecasting accuracy, eliminates pay parity anomalies across agencies, and optimizes municipal budget allocations. Overestimating salaries leads to inflated budget reserves and fiscal waste, while underestimating leads to retention risks, labor disputes, and compensation non-compliance.

ML Benchmark Narrative

On this high-cardinality categorical dataset, gradient boosted decision trees like XGBoost and LightGBM typically achieve robust R² scores around 0.85-0.92 after extensive target encoding of position titles and divisions. In contrast, zero-shot tabular foundation models such as TabICL on Carla HQ infer contextual text embeddings and tabular patterns directly without manual feature preprocessing, delivering competitive out-of-the-box regression performance.

Source Origin:Montgomery County Open Data Portal / OpenML
Creator:Montgomery County Government, Maryland (2016)
License:Public Domain / Open Government Data
FEATURE SPECIFICATION & DATA DICTIONARY

Target & Feature Column Definitions

Exhaustive business definitions, measurement units, roles, and target variables across all 13 columns.

ColVariable / Feature NameRoleTypeUnit / ScaleDescription & Business Meaning
Acurrent_annual_salary_usdCurrent Annual SalaryTARGETNUMUSD
Target variable representing the standardized base annual salary contracted for the employee in USD.
Bfull_nameEmployee Full NameFeatureCAT
Full legal name of the county employee in 'Last, First Middle' format.
CgenderGenderFeatureCAT
Recorded biological sex or administrative gender of the employee ('F' for Female, 'M' for Male).
Dgross_pay_received_2016_usd2016 Gross Pay ReceivedLEAKAGENUMUSD
Total gross compensation paid to the employee during calendar year 2016, including base salary, overtime, and bonuses.
TARGET LEAKAGE RISK(proxy target)

Gross pay received during calendar year 2016 is an ex-post realization of compensation that directly incorporates base annual salary along with overtime and bonuses, serving as an immediate proxy for the target annual salary.

• Excluded during model training & zero-shot inference to avoid artificially inflated metric scores.
Ddepartment_codeDepartment CodeFeatureCAT
Standardized municipal department acronym or abbreviation (e.g., 'POL' for Police, 'HHS' for Health & Human Services, 'COR' for Correction & Rehabilitation).
Eovertime_pay_2016_usd2016 Overtime PayLEAKAGENUMUSD
Cumulative compensation paid specifically for overtime hours worked during calendar year 2016.
TARGET LEAKAGE RISK(post event)

Total overtime pay received is a post-event metric accumulated and recorded over the course of calendar year 2016, reflecting realized historical pay rather than pre-event base contracted compensation.

• Excluded during model training & zero-shot inference to avoid artificially inflated metric scores.
Edepartment_nameDepartment NameFeatureCAT
Full formal name of the municipal operating agency or government department.
Fdivision_nameDivision / SectionFeatureCAT
Specific internal operational division, bureau, or administrative unit within the department.
Gassignment_categoryEmployment CategoryFeatureCAT
Employment status classification indicating full-time vs. part-time and regular vs. temporary status (e.g., 'Fulltime-Regular').
Hposition_titlePosition TitleFeatureCAT
Formal job title and grade level held by the employee (e.g., 'Master Police Officer', 'Social Worker IV').
Iunderfilled_job_titleUnderfilled Job TitleFeatureCAT
Lower classification title assigned when an employee is performing duties of a higher grade position while in training or trainee status.
Jdate_first_hiredOriginal Hire DateFeatureCAT
Calendar date on which the employee was first hired into county service in MM/DD/YYYY format.
Kyear_first_hiredFirst Hired YearFeatureNUMyears
Four-digit calendar year corresponding to the employee's initial appointment date, proxying organizational seniority.
DATASET PREVIEW

Interactive Data Table

Explore rows, feature values, and target labels for Employee Salaries.

Loading dataset...
LOCAL EXECUTION TELEMETRY

TabICLv2 WebGPU Benchmark Results

Complete performance metrics on persisted IID or non-IID evaluation splits, executed 100% locally in the browser sandbox.

Mean Split Latency?Mean wall-clock execution time per persisted evaluation split running 100% locally via WebGPU inside Chrome.Explore Latency Guide →
36.28s9 splits • 8 ensembles/split
Total Evaluation Runtime?Total wall-clock duration across all selected splits, including in-context encoding, WebGPU shader execution, and result aggregation.Explore Split Protocol →
326.49sIncludes warmup & sync
Evaluation Protocol?Precomputed row indices ensure the browser and Python runners evaluate identical out-of-sample observations without leakage.Explore Split Protocol →
3×3 SplitsRepeated IID • 9 total
TabICLv2 Score?Primary task performance score (RMSE) achieved by TabICLv2 across out-of-sample evaluation splits.Explore Metric Formula →
8,126RMSE
Evaluation SplitTrain RowsTest Rows
Split Duration?Wall-clock inference time taken for this evaluation split running 100% locally via WebGPU.Learn more →
R² Score?Coefficient of determination measuring variance explained by the model.Learn more →
MAE?Mean Absolute Error in original target units, robust to outliers.Learn more →
RMSE?Root Mean Squared Error, penalizing large residual deviations.Learn more →
MAPE?Mean Absolute Percentage Error relative to true target values.Learn more →
RMSLE?Root Mean Squared Logarithmic Error, measuring proportional error scales.Learn more →
R1 / F16,1523,07636.77s0.91494,526.838,584.1
R1 / F26,1523,07636.22s0.92044,244.488,244.49
R1 / F36,1523,07636.35s0.91534,679.878,351.42
R2 / F16,1523,07636.19s0.92644,177.777,762.72
R2 / F26,1523,07636.13s0.92784,210.297,920.39
R2 / F36,1523,07636.23s0.92124,203.378,210.76
R3 / F16,1523,07636.11s0.92734,152.338,053.09
R3 / F26,1523,07636.28s0.91974,512.098,434.66
R3 / F36,1523,07636.21s0.92563,933.067,528.85
Mean ± Std36.28s0.9221 ± 0.00474,293.38,121.2
UNIFIED BENCHMARK LEADERBOARD

Foundation Models vs Traditional ML Baselines

Side-by-side evaluation of our in-browser Carla engine (WebGPU), open-source tabular foundation models, and OpenML baselines.

RankAlgorithm / ModelModel FamilyRuntime
R² Score?Coefficient of determination (R²) measuring proportion of variance explained.Google ML Guide & Details →
MAE?Mean Absolute Error measuring average prediction error magnitude.Learn more →
RMSE?Root Mean Squared Error penalizing large prediction errors.Learn more →
Reference / Repo
#1EXAONE TabularTabular Foundation ModelPyTorch/Python0.93483,872.17,426.9LGAI-Research/EXAONE-Tabular ↗
#2Streaming nanotabiclTabular Foundation ModelPyTorch/Python0.92674,148.77,874.3soda-inria/nanotabicl ↗
#3nanotabicl VanillaTabular Foundation ModelPyTorch/Python0.92214,246.28,108.9soda-inria/nanotabicl ↗
#4Carla Engine (TabICLv2 WebGPU)Tabular Foundation ModelBrowser0.92214,293.38,121.2100% In-Browser
#5Full TabICLv2 (PyTorch Reference)Tabular Foundation ModelPyTorch/Python0.92184,253.58,125.8soda-inria/tabicl ↗
#6TabPFN v3Tabular Foundation ModelPyTorch/Python0.91844,260.18,308PriorLabs/tabpfn ↗
#7Google TabFM v1.0Tabular Foundation ModelPyTorch/Python0.87134,882.710,434.5google-research/tabfm ↗
STATISTICAL DISTRIBUTIONS

Variable Schema & Summary Distributions

Observed numerical ranges, category cardinalities, missing rates, and sample values across 9,228 rows.

ColVariable NameTypeMissingDistinctSummary Stats / DistributionSample Values
Acurrent_annual_salary_usdTARGETNUM03,403[9196, 303091] μ=73390.2 σ=29117.569222.1897392.47104717.28
Bfull_nameCAT09,222Aarhus, Pam J.Aaron, David J.Aaron, Marsha M.
CgenderCAT17 (0.18%)22 cats: M, FFMF
Dgross_pay_received_2016_usdLEAKAGENUM100 (1.08%)8,977[0, 313700.42] μ=79504.0 σ=35103.371225.98103088.48107000.24
Ddepartment_codeCAT03710 cats: POL, HHS, FRS +7 morePOLPOLHHS
Eovertime_pay_2016_usdLEAKAGENUM2,917 (31.61%)6,176[-2222.85, 125244.35] μ=10334.3 σ=12677.5416.13326.191353.32
Edepartment_nameCAT03710 cats: Department of Police, Department of Health and Human Services, Fire and Rescue Services +7 moreDepartment of PoliceDepartment of PoliceDepartment of Health and Human Services
Fdivision_nameCAT0694MSB Information Mgmt and Tech Division Records Management SectionISB Major Crimes Division Fugitive SectionAdult Protective and Case Management Services
Gassignment_categoryCAT022 cats: Fulltime-Regular, Parttime-RegularFulltime-RegularFulltime-RegularFulltime-Regular
Hposition_titleCAT0385Office Services CoordinatorMaster Police OfficerSocial Worker IV
Iunderfilled_job_titleCAT8,135 (88.16%)84Firefighter/Rescuer I (Recruit)Traffic Management Technician IPublic Safety Communications Specialist I
Jdate_first_hiredCAT02,26409/22/198609/12/198811/19/1989
Kyear_first_hiredNUM051[1965, 2016] μ=2003.6 σ=9.3198619881989
FREQUENTLY ASKED QUESTIONS

Frequently Asked Questions: Employee Salaries

Common questions regarding the Employee Salaries dataset, machine learning task formulations, and in-browser tabular inference.

What is the Employee Salaries dataset used for?

The Employee Salaries dataset (OpenML ID 42125) is an authoritative public sector tabular benchmark used to evaluate regression algorithms predicting current annual base salaries from department, job title, and tenure metrics.

Can I run zero-server predictions on this dataset in Google Sheets?

Yes, Carla HQ enables zero-server in-context tabular prediction directly in spreadsheets using TabICL foundation models without requiring local Python environments or complex ML pipelines.

What machine learning models perform best on Employee Salaries?

Gradient Boosted Trees (LightGBM, XGBoost, CatBoost) perform well when high-cardinality titles and divisions are properly encoded. Zero-shot foundation models like TabICL achieve comparable accuracy instantly through in-context learning.

LIVE EVALUATION IN GOOGLE SHEETS

Test TabICLv2 on Employee Salaries Yourself

Open the pre-loaded Google Sheet and let Carla configure the target and task for local, zero-cloud tabular machine learning.

Step 1

Launch Carla

Click once to open the spreadsheet and Carla side panel together.

Step 2

Review the Setup

Carla selects the dataset target and task from this page automatically.

Step 3

Evaluate & Predict

Run predictions and compute metrics with zero server uploads.