experiments.


| small & large personal projects.   |   collaborative #dotastro projects.   |   successes & failures celebrated equally.

 
 

#dotastro hack ideas graveyard.

Description:

Need a new hack idea? Bring one back from the dead. The Hack Ideas Graveyard shows forgotten or failed hacks from previous .Astronomy conferences

MOTIVATION:

The .Astronomy Hacks Collector archives projects that were attempted, but not necessarily completed on hack day. But what about the hack ideas that never amounted to anything? Will Armentrout pitched the Hack Ideas Graveyard at .Astronomy9 as a way of preserving the many great (and often silly) ideas. 

TOOLS & DATA:

This responsive website was created using a simple HTML5 template from templated.co. It's hosted on GitHub as a GitHub Page. The data was collected by searching through the #dotastro twitter feed and looking at tweets around the time of each conference, and reading the various .Astronomy live blogs and conference reviews. The website was styled using HTML5 + CSS. Javascript libraries and functions randomise each hack idea (the data) which is contained in a simple .csv file (exported from a shared Google spreadsheet).

TIME TO MAKE / CURRENT STATUS:

This was created in less than a day. There are some design issues that still need to be fixed (image overlays etc). 

CREATORS:

Will Armentrout (@WillArmentrout) – project idea, data collection, and cleaning
Arna Karick (@drarnakarick) – website development/front-end design
Lisa Ballard (@BasilLeaf) – implementing the javascript

badge-img


 

creating color palettes from images.

Description:

A series of Jupyter Notebook tutorials demonstrating how you can create colour palettes from any input image.  The tutorials are based on the Urban Goggles project created by Adrian–Price Whelan (Princeton), Dan Foreman–Mackey (University of Washington), and Ben Nelson (CIERA - Northwestern University) at AstroHackWeek 2016

MOTIVATION:

A visually creative way to explore Python's machine learning package scikit–learn package; specifically the K-means clustering package. I also wanted to turn code from a hack day project into a tutorial that others could learn from. The idea was to go through each step of the process, dissecting some of the functions along the way and adding links to resources, further reading and similar examples.

TOOLS & DATA:

The Jupyter Notebooks are hosted on GitHub.  The Anaconda Python installation was installed prior to the project and the main Python packages used are matplotlib, numpy, and scikit-learn. The images were downloaded from the Apollo Project Archive Flickr galleries and the Planet Labs Gallery.

TIME TO MAKE / CURRENT STATUS:

This was created over two days. The tutorial is essentially complete, but I would like to explore the idea of finding the optimal number of clusters. I would also like to compare the results to colormind.io – an AI powered colour generator that uses deep learning.

CREATOR:

Arna Karick (@drarnakarick)


#dotastro image gallery.

Description:

The .Astronomy Hacks – Image Gallery is a visual feast of hack day projects from the past eight years or so of .Astronomy conferences. It contains most of the hack day projects (those with public websites) from the .Astronomy Hacks Collector. The Hacks Collector contains the full list of hacks, and was conceived by @eteq@DainaBouquin@astrofrog@becky1505@zemogle and @drarnakarick at .Astronomy 8.

MOTIVATION:

The idea for the project came shortly after the .Astronomy8 conference. I wanted to create a more visual version of the .Astronomy Hacks Collector. I also wanted to create a simple a website using the HTML5Up! templates. 

TOOLS & DATA:

The website was developed using the single page HTML5UP!  Multiverse template, and is hosted on GitHub (as a GitHub page)

TIME TO MAKE / CURRENT STATUS:

This was created in a day. The website is  complete up to . Astronomy8

CREATOR:

Arna Karick (@drarnakarick)

 


 
builder_cards_roboph_front.jpg

project builder cards.

Description:

techsavvyastronomer.io project builder cards that showcase what one could build given a set of tools. The cards were created by dissecting .Astronomy hack projects.

MOTIVATION:

I wanted to learn how to create marketing postcards (similar to those found in cafes) using the Canva graphic design platform. I also wanted a simple way to show how you could combine a small number of tools to create innovative projects. The target audience would be novice hackers.

TOOLS:

The logo was designed using GIMP (GNU Image Manipulation) software with icons purchased from Iconfinder.
The cards were designed in Canva. HTML color matching was done using the Color Picker app from the iTunes store.

TIME TO MAKE / CURRENT STATUS:

Four project builder cards were created in a day.

CREATOR:

Arna Karick (@drarnakarick)

 


insight data fellows: word matrix.

Description:

This matrix diagram visualises the research backgrounds of Insight Data Science Fellows and the frequency of fellows represented by that particular combination of research discipline and university.

MOTIVATION:

I wanted to learn how Mike Bostock's co-occurance matrix worked. In hindsight this is probably not the most suitable type of visualisation, although it does convey what I wanted it to.

TOOLS & DATA:

Created using d3.js. Based upon Mike Bostock's excellent Les Misérables co-occurrence matrix example, with modifications taken from Robert Simpson's (Orbiting Frog) data mining blog post .

TIME TO MAKE / CURRENT STATUS:

This was created in a day. The plot contains information about Insight Data Science Fellows up to March 2016. There are no plans to update this with more recent information.

CREATOR:

Arna Karick (@drarnakarick)