Redshift Wrangler

General Citizen Science Questions:

What is citizen science?

Citizen science, sometimes referred to as community science, participatory science, or volunteer science, is a form of scientific collaboration and communication between scientists and volunteers. This use of the term citizen is inspired by the American Astronomical Society, "Every astronomer is a citizen of the community of science", and by the fact that our collaborators are citizens of Earth. Redshift Wrangler requires no prior experience, and we welcome participants of any level of scientific background. Our citizen science project will help you learn more about our research while training you to look for features of interest in our data.

What is the goal of this project?

Our goals for Redshift Wrangler range from our overarching scientific objectives to our plans for open communication and our hopes of including and learning from a broad range of participants.

Collaborating with citizen scientists to identify spectral features and check spectral fits will greatly improve the efficiency of these identifications and provide training sets for eventual automated algorithms. Volunteers going through our huge quantities of data will help us work toward scientific goals like: mapping the cosmic web of distant galaxies, identifying galaxy interactions, finding distant galaxies, calibrating measurements of redshift or distance in space, and training future machine learning algorithms.

Working through each individual spectrum would be too time consuming for a single person, but together with the the help of citizen scientists like you, we can accomplish more science and work through our data faster. Bringing in citizen scientists also lets us review individual classifications, receive feedback from different scientists and non-scientists, get different perspectives on tough data and fits, and provide more access to becoming involved in research for participants at all levels of learning.

Aren't there methods to automate these tasks?

There are! But in their current state, these automated methods are not good enough to identify spectral features and measure the redshifts of galaxies. These typically only work well for the brightest objects with many features. For the vast majority of sources, human visual inspection is still necessary. That's why we need your help! Once we have a redshift, automated routines can fit models to the spectra, but humans still need to inspect the fits to identify any problems. Having more eyes on these measurements will help us work through large datasets more efficiently, make serendipitous discoveries of anything unusual, and help us create training sets for future automation. In this way, automated methods and citizen science will complement each other.

I am not an astronomer and have not studied astronomy; can I still do the classifications?

Yes, of course! Our project is designed for people who are not experts in astronomy or galaxies to be able to participate, and we hope citizen scientists at all levels of prior experience will get something out of Redshift Wrangler.

How can I distinguish what is a real emission/absorption line?

If you look at the overall spectrum and a peak stands out compared to the rest of the data and appears to be more than one pixel/data point wide, then it is most likely a real feature. In Task 1, you can look at both the 1D and 2D data to see if a possible emission line peak shows up as a bright spot in the 2D spectrum - if it does, it's real! If the peak in the panel does not stand out, blends in too much with the rest of the data, looks like blip of just one pixel, then what you're looking at is most likely not a real feature.

In Task 2, if the fit line does not match up with the peak or the general trend of the spectrum then that is also an indicator that it is not a real feature. On the other hand, if the red fit line does match up with a significant peak, then it is most likely a real emission line. You can check out more examples of real features in the field guide for help and review.

How do I know if I am doing this right? What if I made a mistake?

If you are unsure if you are doing it correctly, check the field guide for more good and bad examples of spectral fits and features, and review the tutorials for each task. If you are still unsure if you are identifying data correctly or if you think you made a mistake, don't worry! We are looking at the overall response to each of the data sets. Often the answer may be unclear, and seeing varied responses from participants in these cases is just as useful as seeing cases where the answer if very obvious and unanimous. Like most things, it will get easier with practice, so the more spectra you look at, the more familiar you will become with identifying features.

Am I able to see other responses?

Yes, underneath the "collect" tab you should see other citizen scientist's responses.

---

Task Specific Questions:

What if I can't find emission or absorption features or if there are no real emission/absorption lines?

Not seeing anything is totally fine and you are free to move onto the next object! Often, these galaxies are very faint and it can be hard to separate real features from the background noise. Our goal is to find these features with your help and also make note of objects that are simply too faint to detect.

In Task 1, what should I do if there are bright spots in the 2D that don't correspond to peaks in the 1D?

You may see spectra like this when we have two really close galaxies - check out the Field Guide entry for serendipitous sources to see examples. The 1D spectrum might not correspond to the brightest 2D spectrum, but you can sometimes still spot fainter bright spots below emission lines. We are interested in the 1D spectrum only, so you can ignore any secondary objects that don't correspond to the 1D spectrum.

How do I draw the boxes in Task 1? I'm not able to draw anything on the images.

Make sure you clicked the "Next" button on the task! When you first start the task you need to click "Next" to activate the drawing tool before you're able to draw any boxes. If you still have problems, please reach out to us and we can try and help.

What if I see spectral features in Task 1 AND there's something wrong? Which option should I select?

We want to prioritize marking the spectral features, so if you're still able to mark the emission and absorption lines then you should! If there's something really wrong, you can post it in the talk boards and report it to us.

What should I do if the 2D data in one of the panels looks different than it does in the full spectrum above?

We were finding this issue in our DEIMOS dataset - in a lot of the spectra, there is something wrong with the 2D spectrum in panel f. If you think there might be an emission or absorption line but the 2D data in that panel looks wrong, try looking at the 2D in the full spectrum above. If you're still unsure or think there's something really wrong, you can post it in the talk boards and report it to us. Our team has sorted this out and we don't expect to see this issue in any other datasets.

Why is the data in the new datasets so much noisier?

You may have noticed that our new datasets are not as clean as the original Keck/DEIMOS (DEep Imaging Multi-Object Spectrograph) spectra you initially worked with (and many examples in our tutorials and field guide). That is to be expected for a lower resolution instrument, like GMOS (Gemini Multi-Object Spectrograph) or FMOS (Fiber Multi-Object Spectrograph), which the most recent objects are from. The newest data in Redshift Wrangler also has a higher redshift range; this range can help us detect important rest-frame optical emission lines, but it does mean there may be more atmosphere in the way of our observations that makes the data a lot noisier!