Physical Sciences Datasets


Exoplanet Hunting in Deep Space Kepler labelled time series data
$0.00
The data describe the change in flux (light intensity) of several thousand stars. Each star has a binary label of 2 or 1. 2 indicated that that the star is confirmed to have at least one exoplanet in orbit; some observations are in fact multi-planet systems. As you can imagine, planets themselves do not emit light, but the stars that they orbit do. If said star is watched over several months or years, there may be a regular 'dimming' of the flux (the light intensity). This is evidence that there may be an orbiting body around the star; such a star could be considered to be a 'candidate' system. Further study of our candidate system, for example by a satellite that captures light at a different wavelength, could solidify the belief that the candidate can in fact be 'confirmed'. Description Trainset: 5087 rows or observations. 3198 columns or features. Column 1 is the label vector. Columns 2 - 3198 are the flux values over time. 37 confirmed exoplanet-stars and 5050 non-exoplanet-stars. Testset: 570 rows or observations. 3198 columns or features. Column 1 is the label vector. Columns 2 - 3198 are the flux values over time. 5 confirmed exoplanet-stars and 565 non-exoplanet-stars. Acknowledgements The data presented here are cleaned and are derived from observations made by the NASA Kepler space telescope. The Mission is ongoing - for instance data from Campaign 12 was released on 8th March 2017. Over 99% of this dataset originates from Campaign 3. To boost the number of exoplanet-stars in the dataset, confirmed exoplanets from other campaigns were also included. To be clear, all observations from Campaign 3 are included. And in addition to this, confirmed exoplanet-stars from other campaigns are also included. The datasets were prepared late-summer 2016. Campaign 3 was used because 'it was felt' that this Campaign is unlikely to contain any undiscovered (i.e. wrongly labelled) exoplanets. NASA open-sources the original Kepler Mission data and it is hosted at the Mikulski Archive. After being beamed down to Earth, NASA applies de-noising algorithms to remove artefacts generated by the telescope. The data - in the .fits format - is stored online. And with the help of a seasoned astrophysicist, anyone with an internet connection can embark on a search to find and retrieve the datafiles from the Archive.

Category: Physical Sciences

Keywords: space,astronomy,medium

Rows: 5359

Sales: 0

Questions: 0

Meteorite Landings Data on over 45k meteorites that have struck Earth
$0.00
The Meteoritical Society collects data on meteorites that have fallen to Earth from outer space. This dataset includes the location, mass, composition, and fall year for over 45,000 meteorites that have struck our planet. Notes on missing or incorrect data points: a few entries here contain date information that was incorrectly parsed into the NASA database. As a spot check: any date that is before 860 CE or after 2016 are incorrect; these should actually be BCE years. There may be other errors and we are looking for a way to identify them. a few entries have latitude and longitude of 0N/0E (off the western coast of Africa, where it would be quite difficult to recover meteorites). Many of these were actually discovered in Antarctica, but exact coordinates were not given. 0N/0E locations should probably be treated as NA. The starter kernel for this dataset has a quick way to filter out these observations using dplyr in R, provided here for convenience: meteorites.geo <- meteorites.all %>% filter(year>=860 & year<=2016) %>% # filter out weird years filter(reclong<=180 & reclong>=-180 & (reclat!=0 | reclong!=0)) # filter out weird locations The Data Note that a few column names start with "rec" (e.g., recclass, reclat, reclon). These are the recommended values of these variables, according to The Meteoritical Society. In some cases, there were historical reclassification of a meteorite, or small changes in the data on where it was recovered; this dataset gives the currently recommended values. The dataset contains the following variables: name: the name of the meteorite (typically a location, often modified with a number, year, composition, etc) id: a unique identifier for the meteorite nametype: one of: -- valid: a typical meteorite -- relict: a meteorite that has been highly degraded by weather on Earth recclass: the class of the meteorite; one of a large number of classes based on physical, chemical, and other characteristics (see the Wikipedia article on meteorite classification for a primer) mass: the mass of the meteorite, in grams fall: whether the meteorite was seen falling, or was discovered after its impact; one of: -- Fell: the meteorite's fall was observed -- Found: the meteorite's fall was not observed year: the year the meteorite fell, or the year it was found (depending on the value of fell) reclat: the latitude of the meteorite's landing reclong: the longitude of the meteorite's landing GeoLocation: a parentheses-enclose, comma-separated tuple that combines reclat and reclong What can we do with this data? Here are a couple of thoughts on questions to ask and ways to look at this data: how does the geographical distribution of observed falls differ from that of found meteorites? -- this would be great overlaid on a cartogram or alongside a high-resolution population density map are there any geographical differences or differences over time in the class of meteorites that have fallen to Earth? Acknowledgements This dataset was downloaded from NASA's Data Portal, and is based on The Meteoritical Society's Meteoritical Bulletin Database (this latter database provides additional information such as meteorite images, links to primary sources, etc.)   CREDIT: NASA at kaggle.

Category: Physical Sciences

Keywords: NASA,Government,education,meteor

Rows: 45567

Sales: 0

Questions: 0