VEDA is a desktop laboratory for planetary atmosphere and ionosphere data. Pick a planet or moon and a date range, and VEDA finds every product the official archives hold for it across all connected missions. You can then download the products, plot them, derive physical parameters, compare observations and compute the observation geometry with SPICE.
All data shown in VEDA come straight from the mission archives (NASA PDS, ESA PSA, JAXA DARTS, ISRO ISSDC, OPUS) and the instrument teams' research data repositories (Zenodo, BIRA-IASB). Nothing is simulated. Nearly every payload of every mission in VEDA can be searched by date and plotted; the few that cannot (no searchable archive route yet) are listed in DATA_POLICY.md.
VEDA is public and updated continuously (version 0.2.0; changes go to the main branch continuously, and a new build is released about once a month). It is ready for use and for testing; please report problems, wrong results and suggestions through Feedback in the app or on the issue tracker.
- User guide: searching, plotting, comparison, geometry, export, settings and troubleshooting.
- Data policy and citations: the archives and data sets VEDA reads, their terms, and what to cite.
- Terms of use: warranty, responsibility for results, archive etiquette and privacy.
- License (MIT) and third-party licenses.
- Contributing and changelog.
| Group composites by latitude, light theme | Mission view and archive data | Settings |
|---|---|---|
 |
 |
 |
In By Celestial Body, Find observations of ... searches every connected data set of every mission for that body over the dates you give. Two kinds of data set are searched together:
- Indexed data sets: VEDA reads the archive's own catalogue (PDS3 volume and cumulative indexes, PDS4 bundles) into a local SQLite index the first time, so later searches are instant and work offline.
- Live data sets: whole instrument archives too large to copy are searched on the archive servers for your dates (ESA PSA EPN-TAP, the NASA PDS Registry API, the PDS Rings node OPUS). What is found is kept in the local catalogue.
For example, Mars on 2016-01-01 lists MAVEN NGIMS and IUVS, MRO CRISM, SHARAD and MCS products and Mars Express ASPERA-3, SPICAM, PFS, OMEGA and MARSIS products in one table. Show narrows the list to profiles, images or time series. Open any row to plot it, tick several profiles to compare them, or download them for offline work.
In By Planetary Mission, Archive data lists every payload and data set of the mission (instrument, level, archive, live or indexed). Choose the ones you want and filter by date, product type, free text, profiles only or downloaded only, oldest or newest first.
All 25 missions in VEDA are connected (161 data sets: 55 read from archive indexes, 106 searched live). Highlights:
| Body | Mission | Payloads | Route |
|---|---|---|---|
| Venus | Akatsuki | RS (L2-L4), UVI, IR1, IR2 and LIR (raw, calibrated, geometry); UVI and LIR to the end of the mission (March 2024) with the 2023 LIR recalibration and L3d maps (PDS4) | JAXA DARTS, indexed |
| Venus | Venus Express | VeRa temperature profiles (2006-2009 and 2014, research data on Zenodo) and SPICAV-SOIR profiles (2006-2014, BIRA-IASB); VeRa raw data, VMC, VIRTIS, SPICAV/SOIR, ASPERA-4, MAG | ESA PSA, live; Zenodo; BIRA-IASB |
| Venus | Magellan, Pioneer Venus, VEGA | Radio occultation profiles (Magellan; PVO 1978-1989 temperature-pressure and electron density, recovered by Withers et al. 2020), radar, gravity; the VEGA 2 lander descent (63 km to the surface) and VEGA balloon records; PVO OIMS, ONMS, probes | NASA PDS |
| Mars | Mars Express | MaRS L4 profiles (indexed); SPICAM, PFS, OMEGA, ASPERA-3, MARSIS, HRSC, VMC | ESA PSA |
| Mars | MRO | MCS DDR/EDR/RDR (indexed), CRISM, SHARAD; CTX and MARCI on request | NASA PDS |
| Mars | MAVEN, MGS, MRO, Mars Odyssey, MER, Phoenix, MSL, InSight | NGIMS, IUVS, ACC; MGS and MRO radio occultation profiles; MRO and Mars Odyssey aerobraking density profiles; MRO CRISM dust and water-ice aerosol profiles; entry profiles of Spirit, Opportunity, Phoenix, Curiosity and InSight; TES, MOLA | NASA PDS |
| Jupiter | Juno, Galileo | MWR, JIRAM, UVS, JunoCam, MAG, gravity; Galileo SSI, NIMS, UVS, PPR, MAG, PLS, EPD, probe | NASA PDS, OPUS |
| Saturn, Titan | Cassini-Huygens | ISS, VIMS, UVIS, CIRS (OPUS), INMS; RSS profiles of the Titan neutral atmosphere and ionosphere and of the Saturn ionosphere; UVIS Saturn thermosphere profiles; Huygens HASI, DISR, GCMS, ACP, DWE, SSP | NASA PDS, OPUS, ESA PSA |
| Pluto | New Horizons | REX, Alice, LORRI, Ralph MVIC/LEISA, SWAP, PEPSSI, SDC | NASA PDS |
| Mercury | MESSENGER, BepiColombo | MLA, GRS, NS, XRS, MAG, MASCS, RS; BepiColombo MPO-MAG, MCAM, SERENA, PHEBUS, MERTIS, MORE ... | NASA PDS, ESA PSA |
| Moon | LRO | Diviner, LOLA, LEND, Mini-RF, radio science | NASA PDS |
| Ceres, Vesta | Dawn | GRaND, gravity | NASA PDS |
| 67P | Rosetta | OSIRIS, VIRTIS, ALICE, MIRO, ROSINA, RPC, RSI, GIADA, COSIMA and the lander instruments | ESA PSA |
| Mars, Moon | MOM, Chandrayaan-2 | All payloads, via sign-in and import (ISRO PRADAN requires an account) | ISRO ISSDC |
DATA_POLICY.md lists every data set. ISRO's PRADAN portal requires a free account, so VEDA opens it for you to sign in; you download there and use Import downloaded files. VEDA never sees your password.
Any product VEDA lists can be opened: Open for atmosphere profiles, View for everything else. The product viewer reads PDS3, PDS4, FITS and netCDF:
- Tables and time series: plot any field against any other (time axes, one plot or one panel per field, log scales). Large products are drawn at a few thousand points with every minimum and maximum kept.
- Spectra per row (energy channels, spectrometer frames): spectrograms with the mean spectrum.
- Profiles per row (e.g. MRO MCS, about 370 temperature, pressure, dust and ice profiles per file): plot one vector against another for any row, and overlay up to 12.
- Images and maps: stretch (percentile, ZScale, linear, log, sqrt, asinh, histogram), colour maps, bands, value read-out, line transects and histograms; netCDF maps (e.g. Akatsuki L3) keep their longitude-latitude axes.
- Spectral cubes (VIRTIS, OMEGA, VIMS, CRISM, IUVS ...): band slider and the spectrum at any clicked pixel.
Binary data are memory-mapped and images are read at screen resolution, so very large files do not fill your memory.
An opened profile shows every quantity in the product with the archived 1-sigma uncertainty where the product gives one, and its time, latitude, longitude, solar zenith angle and local time. Units are read from the label and converted (Pa, hPa, bar, mbar; K and degrees C; m^-3 and cm^-3, including scaled units). From temperature and pressure VEDA derives, with the body's own constants (
- Lapse rate
$\Gamma = -dT/dz$ and gravity$g(z) = g_0 (R_p/(R_p+z))^2$ - Scale height
$H = R_{spec} T / g$ and speed of sound - Potential temperature
$\theta = T (P_0/P)^{R_{spec}/c_p}$ and mass density$\rho = P/(R_{spec} T)$ - Brunt-Vaisala frequency
$N^2 = \frac{g}{T}\left(\frac{dT}{dz} + \frac{g}{c_p}\right)$ and buoyancy period - Cold-point tropopause (Titan, Jupiter, Saturn), gravity-wave temperature perturbations and potential energy
- For ionospheres, the peak height and density, a fitted Chapman layer and
$\text{VTEC} = 10^{-7}\int N_e,dz$ (TECU) - From a measured density (accelerometer and entry-probe profiles, SOIR, radio occultation number densities): temperature and pressure by downward hydrostatic integration
- For profiles with temperature and pressure: their hydrostatic consistency (largest and median departure of the archived pressure from the hydrostatic one)
Heat capacity depends on temperature for Venus, Mars, Titan and Pluto (
- Choose the profiles by date, latitude band, local solar time, solar zenith angle, Mars season (Ls) and number per mission: VEDA searches the whole archive catalogue, downloads what it needs and reports why profiles were left out. Or tick profiles in any search table.
- Common grid: profiles from any missions are interpolated onto common altitude levels (grid step of your choice) or pressure levels (uniform in log pressure, so profiles with different altitude references line up), with no extrapolation and no bridging of data gaps, and drawn with their mean and 1-sigma spread: arithmetic for temperature-like quantities, geometric with a multiplicative spread for pressure, densities and electron density.
- Climatologies: group composites by latitude band, local time, zenith angle, Ls, year, month, month of the year or mission, and deviations from the group or composite mean.
- Altitude cut: one point per profile against time, latitude, local time, zenith angle, Ls, longitude or day of year: the variable at one altitude, its minimum, maximum or mean in a layer (or their altitudes), or a quantity from the whole profile (tropopause, electron density peak, Chapman fit, electron content, gravity-wave energy, hydrostatic consistency). Against local time, longitude or Ls a least-squares fit of up to four harmonics gives tide and wave amplitudes and phases with 1-sigma uncertainties.
- Colour the curves by mission, date or latitude and switch the variable and units at any time.
- Uncertainties: the archived 1-sigma values are carried into every derived quantity (first-order formulas, Monte Carlo for derivatives and the temperature retrieved from density) and drawn as bands; composites have standard errors, 95 % bootstrap intervals, optional inverse-variance weighting and the effective number of profiles; tide fits and the points of the altitude cut have bootstrap intervals.
- Filters and outliers: profiles can be required to cover an altitude range, to have a fine enough level spacing or a small enough uncertainty, and can be drawn from chosen data sets or instruments; an outlier screen (robust z per level) flags profiles far from the others and leaves them out only when asked.
- Analysis: a zonal-mean latitude-altitude cross section with standard errors, vertical wavenumber spectra of temperature perturbations with their slope, correlation and regression between any two profile quantities with confidence intervals, and on Venus the Venus-GRAM 2021 / VIRA reference atmosphere.
- Comparison CSV: the compared variable on the common grid (composite and group means, spreads, standard errors and bootstrap intervals, profiles and effective profiles per level, every profile with its 1-sigma) with a header line per profile (mission, time, geometry, altitude reference) and a recipe line from which Open comparison redoes the same comparison.
- Profiles, all variables (CSV): the compared profiles at their own levels with every archived and derived quantity and uncertainty, one row per profile and level, ready for pandas, R or a spreadsheet.
- For an opened profile: CSV and structured JSON with its derived quantities; every CSV names the archive, source file, citation and VEDA version.
Plot style controls lines, markers, palettes, uncertainty bands or error bars, linear or log axes, swapped axes, altitude or pressure as the vertical axis, grid, ticks, fonts and legend, with journal templates for AGU, Elsevier (Icarus/PSS), A&A and MNRAS. It applies to profiles, comparisons, time series, spectra and images. Export figure writes PNG at a chosen DPI or vector SVG at the journal's single- or double-column width.
Geometry is set up for 17 of the 25 missions (not yet for Mars Odyssey, the Mars landers and rovers or VEGA; MOM and Chandrayaan-2 publish no kernels):
- Radio occultations: the orbit around the occultation (planet-fixed and J2000), the view from Earth, the tangent-point track (cylindrical, polar or orthographic) and SZA, local time and Sun-Earth-probe angle along the profile. Times are Earth-received and light-time corrected; tangent radii agree with Mars Express's archived values to about 1 km.
- Any other observation: the spacecraft orbit, the sub-spacecraft ground track and altitude, local solar time and solar zenith, emission and phase angles over the observation.
When you open a mission, its generic and body kernels download in the background; when you open an observation, the spacecraft ephemeris for that date follows. VEDA knows each mission's kernel archive (NAIF, ESA SPICE service, JAXA DARTS) and how its files map to dates. Settings > Observation geometry turns this off or sets the size above which VEDA asks first. Kernels are kept and reused.
Cite lists the references for exactly what you used on this computer:
- the data sets (with the archive data set identifiers of the products you opened);
- the instrument and mission papers (checked against Crossref);
- the archive acknowledgements;
- SPICE and SpiceyPy if you used geometry, and NumPy, SciPy, Astropy, Matplotlib or Plotly for the features you used;
- VEDA itself.
It gives BibTeX, plain text and a matching data availability statement, all ready to copy.
Load File or drag and drop: PDS3 (.lbl + data), PDS4 (.xml + data), CSV and text tables, FITS and PNG/JPEG images (up to 200 MB per file). For tables VEDA shows the columns with their units and ranges and asks what each one is and in which unit, and which body, mission, instrument and time the file belongs to; loaded profiles then take part in comparisons like archive profiles.
Open the latest release and download the archive for your OS:
| OS | Asset | Run |
|---|---|---|
| Windows 10/11 (x64) | VEDA-<version>-windows-x86_64.zip |
unzip, open the folder, double-click VEDA.exe (keep the _internal folder next to it) |
| macOS (Apple Silicon) | VEDA-<version>-macos-arm64.zip |
unzip, open the folder, right-click VEDA.app > Open the first time (the app is not notarized) |
| Linux (x86_64) | VEDA-<version>-linux-x86_64.zip |
unzip, cd VEDA-*; chmod +x VEDA && ./VEDA |
Updates. Changes go to VEDA's main branch continuously, tested on all three systems. About once a month they are built and published as a new release ("VEDA 0.2.0 build N"): the release workflow checks every week and publishes when there are new changes and the last release is at least 30 days old. Only the latest and the previous release can be downloaded from GitHub; every release stays archived on Zenodo with its own DOI, under VEDA's concept DOI 10.5281/zenodo.23215291. VEDA checks for a newer build when it starts and shows a notice with a download link; About > Check for updates checks on demand. A published build also updates itself: once a week it looks for the latest release (a copy that missed several, for example one offline for months, goes straight to the newest) and downloads the build for your system in the background (checked against the size and SHA-256 digest GitHub lists for it) and installs it the next time it starts, replacing only its own folder (the previous one is kept as <folder>.previous until the new build has started). Your data, downloads, settings and exports live in a separate folder (below) and are never touched; after an update VEDA shows what is new and any settings that were added or changed, keeping your values. Settings > Network turns the weekly update off and has Check now. To update by hand, download the new archive and replace the old VEDA folder.
pip install "git+https://github.com/jovian-explorer/VEDA.git"
veda # desktop window (falls back to the browser)
veda --browser # always use the default browser
veda-server # headless API + web UI, prints the URL
python -m veda # same as `veda`netCDF files (classic and HDF5-based netCDF4) are read with the netCDF4 package, which is installed with VEDA.
git clone https://github.com/jovian-explorer/VEDA.git
cd VEDA
python -m venv .venv
# Windows: .venv\Scripts\activate macOS/Linux: source .venv/bin/activate
pip install -e ".[dev]"
pytest| Command | What it does |
|---|---|
veda |
Starts the backend on a free local port (8765 or the next free one) and opens the native window |
veda --browser |
Same, but opens your default browser |
veda --no-window --port 8765 |
Server only; open http://127.0.0.1:8765/ yourself |
veda-server --host 0.0.0.0 |
Serve on your network; set VEDA_ALLOWED_HOSTS to the host names or addresses others will use (for example myhost,192.168.1.20). The API has no accounts, so only do this on a trusted network |
The interactive REST API documentation is at /api/docs on the same address. The archive endpoints are under /api/veda/archive (data sets, index, search, fetch, profile) and /api/veda/geometry.
Where VEDA keeps your files (archive index, downloaded products, SPICE kernels, exports, settings, logs):
| OS | Location |
|---|---|
| Windows | %LOCALAPPDATA%\VEDA |
| macOS | ~/Library/Application Support/VEDA |
| Linux | $XDG_DATA_HOME/VEDA (default ~/.local/share/VEDA) |
Set VEDA_HOME to use a different folder. Deleting the cache folder is safe; VEDA downloads again what it needs.
An internet connection is needed to search an archive for the first time, to download products and to fetch SPICE kernels. The bundled sample products (Akatsuki, Mars Express, Cassini) work offline. Turn Settings > Network > Allow downloads off to work fully offline.
pip install -e ".[build]"
python scripts/build_exe.py # dist/VEDA/VEDA.exe, dist/VEDA.app or dist/VEDA/VEDA
python scripts/build_exe.py --archive # also zips it with the licences and guides for distributionPyInstaller cannot cross-compile, so each OS builds its own binary. .github/workflows/release.yml builds all three and publishes them as "VEDA 0.2.0 build N", with the commits since the previous release as release notes, about once a month: a scheduled run every Monday publishes main when it has commits since the latest release, that release is at least 30 days old (repository variable RELEASE_MIN_DAYS) and CI passed on main's head commit. It can also be run by hand (publishes main's head if CI passed on it), and it publishes every pushed version tag (v1.2.3). After publishing, it deletes every release but the latest two from GitHub, keeping their tags (Zenodo's records and the app's links to the changes between releases point at them); each release stays archived on Zenodo with its own DOI. Before publishing, each build's update swap is tested on its own system (scripts/check_update_swap.py).
- Windows: uses the Microsoft Edge WebView2 runtime (preinstalled on Windows 10/11).
- macOS: uses the system WebKit, so nothing extra is needed. Unsigned builds need right-click > Open the first time.
- Linux: the native window needs GTK or Qt bindings for pywebview, for example
sudo apt install python3-gi gir1.2-webkit2-4.1orpip install "pywebview[qt]". Without them VEDA opens in your browser automatically.
VEDA does not own or host any data. Each archive's terms apply to the data you download (see DATA_POLICY.md); the MIT License applies to the VEDA software. When you publish results, cite the data set, the instrument team's reference publication and VEDA. Data & Licenses in the app lists every connected data set with its citation and gives a data availability statement you can copy.
@software{Aggarwal_VEDA_2026,
author = {Keshav Aggarwal},
title = {{VEDA: Visualization, Exploration, and Data Analysis - A Multi-Mission Planetary Science Data Laboratory}},
year = {2026},
version = {0.2.0},
doi = {10.5281/zenodo.23215291},
url = {https://github.com/jovian-explorer/VEDA}
}The DOI is Zenodo's concept DOI for VEDA: it covers every archived release and always resolves to the newest one; each release also has its own version DOI, listed on the Zenodo record.
- License: MIT. Copyright (c) 2026 Keshav Aggarwal.
- Terms of use: TERMS.md. VEDA is provided without warranty, has no accounts, analytics or telemetry, and contacts only the archives you search.
- Third-party software: THIRD_PARTY_LICENSES.md (Plotly.js, KaTeX, FastAPI, Pydantic, SpiceyPy/CSPICE, NumPy, SciPy, Astropy, Matplotlib, Pillow, Requests, Uvicorn, PyWebView).
- VEDA is not affiliated with or endorsed by NASA, ESA, JAXA or ISRO.
src/veda/archives/ archive data sets, catalogue index, downloads, profile loading
src/veda/geometry/ SPICE kernel planning/download and observation geometry
src/veda/readers/ PDS3, PDS4, FITS and text readers
src/veda/analysis/ derived atmospheric and ionospheric parameters
src/veda/api/ FastAPI backend
src/veda/frontend/ web UI (Plotly and KaTeX bundled for offline use)
src/veda/sampledata/ real archive products used by the demos and tests
packaging/ PyInstaller spec, entry script and icon
scripts/ build_exe.py
tests/ pytest suite
pytest # full suite
pytest tests/test_archives.py # a single moduleCI runs the suite on Windows, macOS and Linux for every push to main. The Windows exe launch tests in tests/test_explorer_launch.py run after python scripts/build_exe.py and skip otherwise.
Keshav Aggarwal, Student visitor at the Space Physics Laboratory (SPL), Vikram Sarabhai Space Centre (VSSC), Indian Space Research Organisation (ISRO), Thiruvananthapuram, Kerala, India. Formerly Prime Minister's Research Fellow, Department of Astronomy, Astrophysics and Space Engineering (DAASE), IIT Indore.
Research: planetary radio occultation, planetary atmospheres and ionospheres, solar wind velocity and turbulence, and coronal electron density.
Website: https://jovian-explorer.github.io/