Arclight |

Welcome to the Arclight Guide! This guide will help you get started by introducing you to the basic functions and options Arclight offers.

Performing a Search

Performing a search in Arclight is easy and fast. In the query box in the top left corner, simply enter an query and hit enter. The resulting graph will show how your query trends across the Media History Digital Library. By moving your cursor over the data points, Arclight will tell you how many page hits your query had in that particular year. Additionally, when you click the data point, it will take you to Lantern so you can have a closer look at these specific pages.

Performing a Multiple Query Search

Arclight isn’t limited to one query. If I wanted to search for the cast of Grand Hotel (1932), I can take a list and copy and paste the queries into the search box and hit enter. Note that if you are using a word processor, each query on the list should be on a new line.

When performing searches with several queries, the results may be overwhelming. On the left side of the screen will be tabs with each query. When you move your cursor over a tab, it gives you some options. The large X removes the query from search, while Eye button toggles the visibility of the query on and off.

Setting a Year Range

This button will allow you to set date limits on your searches. When you click this button, a tab will appear where you can enter the years you want to focus on. When the years limits are set, the graph will adjust and scale to your time period.

The Maximum Hit Guide

This button will add the Maximum Hits Guideline to your graph. This guideline tells you the total number of pages the Media History Digital Library has for each year. This guideline reveals some important information, such as gaps in our corpus (e.g. 1901-1902) and years where we have the most pages. With this, we can make better interpretations of our results (see Normalizing Data below for more).

Normalizing Data

This button will take the number of page hits of each year and divides it by the total number of pages available in that year (the Maximum Hits Guideline) to return a percentage. This is helpful because the resulting graph now shows the results on a common scale.

For example, let’s compare two graphs about Greta Garbo. This graph shows the number of pages hits per year during the time period from 1925 to 1935. Notice how Greta Garbo peaks in 1929 with over 800 page hits.

The graph below shows the same search, except the data is normalized. Instead of page hits, the graph refers to percentages. In 1929, Greta Garbo was registered in about 2.3% of the Media History Digital Library‘s pages for that year. Notice how in this graph, her results peak in 1932; while her overall page hits were relatively smaller in 1932, she was in a higher percentage of the available pages. This could be due to several reasons: increasing or waning popularity, changing marketing practices, the varying makeup of the corpus between years, and/or simple mathematical statistics as 1932 has over 10,000 fewer pages than 1929. While this warrants further investigation, the main point is that we should consider larger contexts when looking at our results and normalizing our data helps us with this.

Stack Series

This button will allow you to combine or “stack” queries so that you can see how they trend together. Starting at the front with the last query entered, Arclight takes those numbers and adds it to the next layer behind, and repeats until it reaches the first query entered.

In the graph below, we have a search for Greta Garbo and Joan Crawford (in that order). In the front layer, the pink Joan Crawford, Arclight is giving her basic results. In the orange layer, Arclight is combining both numbers. In this case, Garbo and Crawford had over 1,500 combined hits in 1930.

For more than two entries, the process continues until it reaches the back layer. In the graph below, Yellow is John Barrymore, Pink is Barrymore and Joan Crawford, and Orange is Barrymore, Crawford, and Greta Garbo.

Journal Filter

This button allows you to limit your search to specific journals. Clicking the button will produce a menu that will allow you to search and select which journal(s) you wish to search. Once selected, the journal title will appear on the right side. You can remove journals by clicking the X next to their titles. Once you are done, click outside the menu and Arclight will automatically perform the searches.

Exporting Data

After you have completed your searches, Arclight lets you download the graphs and data in multiple formats. Clicking this button will bring up a menu where you can select your preferred format:

We hope this Introduction Guide has been helpful and if you have any further questions, don’t hesitate to contact us here.

#DHSCMS: Digital Humanities, Tools, and Approaches at SCMS 2015

by Derek Long

April 9, 2015Arclight, conference reports, DHSCMS, SCMS

#DHSCMS: Digital Humanities, Tools, and Approaches at SCMS 2015

This conference report is reposted with the kind permission and cooperation of Antenna.

This year’s Society for Cinema and Media Studies conference in Montreal featured a number of excellent panels that were broadly dedicated to “the digital.” Eric Hoyt (@HoytEric) and I (@DerekLong08) live-tweeted digitally-themed panels using the #DHSCMS hashtag, through our personal Twitter accounts as well as the Project Arclight account (@ProjArclight). We defined the digital broadly in our Twitter coverage, attending any panels that conceptualized the digital as a research tool, a methodological approach, or an object of study. What follows are some themes, questions, and trends that emerged from those panels, along with brief thoughts on that status of digital work in our field, and how it might move forward. This summary is far from exhaustive: there were many compelling presentations beyond the ones discussed here. For more extensive coverage, check out #DHSCMS on your preferred Twitter platform.

Wednesday’s panel on “Network Studies” (B19) situated the digital as itself an object of study, and offered a number of valuable insights into historical and contemporary digital practices. Steven Malčić offered a particularly valuable history of the early Internet, carefully contextualized in object-oriented ontology as well as the history of computer science as a discipline. According to Malčić, ARPANET engineers conceptualized network “entities” in an inverted way, positing the stability of “processes” (protocols and applications) and the ephemerality of “objects” (users, hosts, and servers). Sheila Murphy, in her presentation on fitness trackers and other wearable devices, showed some of the ways in which personal data collection on them has been framed as a game. She made an important distinction between two different models of data collection: that of all-powerful, active “surveillance,” and the more passive “capture model” of wearables. Malčić and Murphy’s papers were only two of the many that explored the important intersections between the digital-as-technology and the digital-as discourse/rhetoric/philosophy, demonstrating that our understanding of how digital technology is implemented is fundamentally linked to the ways in which it is conceptualized.

Thursday offered a pair of workshops on digital archives and their uses, and both testified to the ways that digital technologies are changing research, publishing, and pedagogy. The workshop on “Making the Past Visible” (F19) modeled a number of innovative practices. Michael Newman made a compelling case for the use of archival images as evidence instead of simple “illustration.” He showed how platforms like Pinterest and Tumblr might be used as a supplement to the traditional book publishing model, not only as a means of preserving color, sharing evidence productively, and making visual arguments, but also as a way to connect scholars and other users with common interests. Deborah Jaramillo made the interesting point that pedagogy should be considered an integral part of archival research; teaching students with documents can be a way to curate the often-intimidating volume of digital archival documents. Curation was a theme that returned throughout the workshop, with several participants arguing for a greater valuation in our field—particularly for the purposes of hiring and promotion—of data collection, processing, and sharing.

A pedagogically-focused workshop later in the day (G18) showcased several resources for and approaches to teaching film and broadcast history. Beth Corzo-Durchardt stressed the growing importance of teaching students how to evaluate online sources, as well as how to be collaborative in their primary document research. Catherine Clepper modeled a fascinating assignment using the Cinema Treasures website (cinematreasures.org) to engage students with the exhibition history of their own hometowns, while Eric Hoyt, citing Franco Moretti, argued that the humanities needs to do a better job of teaching students how to read data at scale. Toward that end, Hoyt gave first public demonstration of Arclight, an app that visualizes hits, by year, for search terms across the entire corpus of the Media History Digital Library.

In keeping with Hoyt’s hands-on demonstration of the turn toward digital archives, Martin Johnson made the crucial point that despite this turn, scholars often proceed as if they are still using physical collections. This echoed nicely with the valuation point made in the earlier workshop, hammering home the importance of crediting—and being critical of—digital collections.

Finally, Friday afternoon’s panel on Digital Film Historiography (L8) showcased a large-scale digital project underway at the EYE Film Institute in Amsterdam. Christian Gosvig Olesen presented on the project, which uses the 900 early films, 2000 photos, and 120,000 business documents of the almost completely digitized Jean Desmet collection. As Olesen explained, the Desmet project combines a mapping of early film distribution in the Netherlands (based on the business documents) with ImagePlot visualizations of the collection’s color films, in an exploration of the ways in which New Cinema History and statistical style analysis might be combined. Work toward combining these two very different modes of analysis for the project’s ultimate question—Did films with certain patterns of color aesthetics correlate with particular distribution patterns?—remains in an experimental stage. However, the project’s mapping component offers an interesting model for laying bare the inherent uncertainty and often-invisible lacunae of digitized corpora. By using color, the Desmet project clearly distinguishes between all the films in the collection and those for which distribution data actually exists, modeling one solution to the problem of corpus transparency in digital scholarship.

My personal takeaway from Montreal is that scholars are increasingly using digital tools for every aspect of their work, be it analysis, archival research, publication, or pedagogy—and with a healthy critical understanding of their advantages and pitfalls. Truly exciting work is being done in all of those arenas. What remains, as I see it, is the problem of technical implementation. Aside from a breakout session in the pedagogy workshop, none of the panels I attended offered practical training in the use of software tools or coding. Even a broad discussion of the kinds of tools, platforms, or languages that might be useful for approaching particular problems or research questions would help to expand the base of scholars able to do digital work. Some kind of forum focusing on the use of specific digital tools would almost certainly be well-attended if it were to be offered at next year’s conference, whether as a workshop, practicum, or even a poster-style “drop-in” exhibition. Such a forum would be invaluable, even if it were only a starting point; getting hands-on has a wonderful way of reducing the intimidating character of some more advanced digital tools, for students and advanced scholars alike. If the last few years have seen a digital turn in media studies, then solving the problem of technical implementation may very well usher in a digital acceleration.