Digitizing the Sky

In the 1880s, Edward Charles Pickering, a stout Harvard astronomer whose deeply angled eyebrows recall an angry cartoon character, took on a new project: photographing the entire sky. Using a novel emulsion method, Pickering set about systematically copying narrow slices of night sky onto glass photographic plates.

He planned to use the images to create a catalogue of the stars—he wanted to count them, measure them, and start a classification system.

But as he amassed plates and calculations, Pickering soon found himself with more data than he knew what to do with.

To manage his rapidly growing collection of plates, Pickering took an unorthodox step: he hired a woman. The Harvard Observatory had set a radical precedent only a few years earlier by hiring female staffers, but most of these women worked as observers or assistants.

Pickering offered his hire, Scottish housemaid Williamina Fleming, a much higher-level job: she began working as a part-time computer (someone who “computes”) for the Observatory. Pickering’s initial support helped set Fleming on the path to success: in 1889, she was named the first Curator of Astronomical Photographs and became the first woman to hold an official title at the University.

Throughout his career, Pickering continued to hire women and often promoted them. Until his death in 1919, Pickering typically employed between 12 and 20 women at all times. His unique team of female astronomers came to be known as “Women Computers” or—among those less eager to promote women in science—as “Pickering’s harem.”

Pickering and his team’s work often amounted to a few repetitive chores—making images and physically measuring the diameter of the small circle of light emitted by a few select stars. Astronomical Photographic Plate archivist Lindsay Smith says the women would often use a small tool called a “fly-spanker,” so called because the devices look like shrunken fly-swatters made of glass.

“What it is, is a piece of a plate that has different [star] magnitudes on it,” Smith explains. “They could measure different stars on other plates just by using this tool, because they had marked which magnitude each of those dots represents. They could hold it up to a plate, move quickly, and measure the stars.”

Though the all-female team disbanded in the 1930s, the women computers’ work went on for years—the final photograph was taken in 1992. By that point, the collection comprised more than half a million photographic plates: a visual documentary of a century of sky. Jonathan E. Grindlay first encountered this collection as a graduate student at Harvard in the 1970s. Grindlay soon realized that this century-spanning collection had important implications in a rapidly developing field called “Time-Domain Astronomy,” or the study of how astronomical objects change over time.

Particularly exciting to Grindlay, now a professor of Practical Astronomy and a specialist in Black Holes, was the prospect of using variability in brightness over time to measure supermassive black holes and locate black holes within our own galaxy.

But it was simply impossible to carry out this kind of research manually. There was just too much data to do the mass-scale measurements Grindlay hoped to perform.

“From 1880 to 1992, most of the work involved looking at a miniscule fraction of what’s on all of these plates,” Grindlay says. “What we wanted to do from the beginning was to measure every object on every plate. There could be easily 100,000, maybe 200,000 stars on just [one] plate. They were measuring maybe a dozen objects at a time, but there’s no way you could do it all.”

When Grindlay was a graduate student, computer technology was still new, scanners were slow, and disc space was miniscule. It wasn’t until 2004 that Grindlay got to realize his decades-old dream of digitizing the plates to take stock of every star they held.

Once computers caught up, Grindlay embarked on a mission much like Pickering’s: to digitize the entire sky. Grindlay’s project—“Digital Access to a Sky Century @ Harvard” (DASCH)—now lives in an Observatory building where some of the women computers once worked. The project uses a state-of-the-art scanner to take 60 overlapping images of each slide and stitch them together into a crystal-clear scan.

“We go and we fit coordinates [to the image] that match the coordinates on the sky,” Grindlay says. “Then, we compare every single one of them to existing catalogues, which we now have of the whole sky. We didn’t have that even 20 years ago. And from that catalogue we determine ‘Yes, this is that star in this catalogue,’ or ‘No, it’s a new thing we’ve never seen before.’”

The scanner can process two plates in less than a minute and a half, but even at such speeds, the 13-year-old project is still two to three years away from completion.

Grindlay says when the project is complete, the full century of digital sky will be available for public consumption.

“It’s already open,” Grindlay says. “We’re releasing it as we go. For big projects like this we have data releases, because there’s no way we could look at all this data, you could do so many different things with all this.”

When the last plate is digitized, the DASCH project will have over 1,000 terabytes of data and Pickering’s mission will finally be realized, nearly 150 years after the astronomer first set out to photograph the sky.