Background: Education teaches children about the world, its people, and their place in it. Much of this happens through the curricular materials we present to children, particularly the books they read in school and at home (Giroux, 1981; Apple and Christian-Smith, 1991; Jansen, 1997; Van Kleeck, Stahl and Bauer, 2003; Steele, 2010). How different identities are portrayed in these books has the potential to shape children's beliefs about what roles they and others can or cannot inhabit. Given persistent racial and gender inequality in society (Darity and Mason, 1998; O'Flaherty, 2015; Blau and Kahn, 2017), and the importance of identity and representation in driving beliefs, aspirations, academic effort, and outcomes (Dee, 2005; Riley, 2017; Gershenson et al., 2018; Porter and Serra, 2019), these representations offer a key means through which we either address, perpetuate, or entrench core societal inequalities. Objectives: We use new and established computer vision and natural language processing tools to measure the representation of race, gender, and age in the images and text contained in influential collections of children's books. Our work makes two core contributions. First, we develop and showcase new tools for the systematic analysis of images, highlighting their potential use in a wide range of applications in policy, education practice, and social science research. Second, we apply these tools, alongside established text analysis methods, to compare measures of racial and gender representation across the images and text in collections of children's books and see how these measures have changed over time. Setting: Our study focuses on showcasing these new tools as well as using them to measure representation in books that have won awards from the Association for Library Service to Children, a division of the American Library Association, starting in 1922. Data Collection and Analysis: We digitize the books in our main dataset and then apply frontier computer vision and natural language processing tools to analyze the messages contained in their text and images. Historically, content analysis to measure these messages has been done "by hand" using human coders (Bell, 2001; Neuendorf, 2016; Krippendorff, 2018). Such analysis provides deep understanding but can generally only be done on a small set of content and necessarily reflect human behavior and biases. Our tools allow for standardized, replicable content analysis to be performed on large collections of content such as the one we study. Our first contribution is to showcase new tools for analyzing representation of race, gender, and age in images. Analyzing images involves three primary components: (1) training the computer to detect faces, (2) classifying skin color, and (3) predicting the race, gender, and age of the faces. We build on existing face analysis software tools, making key improvements that allow us to i) analyze the illustrations in children's books accurately (previous models focus on classification of photographs), ii) classify the color of characters' skin, an important dimension of representation which, to the best of our knowledge, was not measurable using previously existing tools, and iii) increase accuracy of existing models. For our second contribution, we apply these tools, alongside established computer-driven text analysis tools, to quantify the levels of race and gender representation these children's books have exhibited in their text and images and how these messages have changed over time. We present our results for two primary groups: (i) "mainstream" books considered to be of high literary value but written without explicit intention to highlight an identity group (Newbery and Caldecott Awards) and (ii) "diversity" books selected because they highlight experiences of specific identity groups (e.g., Coretta Scott King and Rise Feminist Awards). Findings: We present a series of descriptive analyses documenting patterns of representation in the images and text of these books over time. Additionally, we explore the efficacy of explicit efforts to highlight diversity and their likelihood to account for intersectional experiences. We find that books in the Mainstream collection -- which are twice as likely as other children's books to be checked out from libraries -- are more likely to depict lighter-skinned characters than those in the Diversity collection, even when looking at characters within the same race. Particularly surprising is that, despite there not being systematic differences in skin tones across ages in society, children are more likely than adults to be shown with lighter skin in all collections. We also find that the Diversity collection has broader geographic representation of famous figures born outside of the United States or Europe than the Mainstream collection. We compare the incidence of female appearances in images to female mentions in text, and consistently find that females are more likely to be shown in images than discussed in the text, except in the collection of books specifically selected to highlight females. This suggests there may be symbolic inclusion of females in pictures without their substantive inclusion in the actual story. Despite being half of the US population, females have been persistently more absent than males -- especially White males -- on average, in both images and text. This is regardless of the measure used: predicted gender of the pictured character, pronoun counts, specific gendered words, famous figure gender, character first names, and geographic origin. Conclusion: Our tools will directly contribute to lasting improvement of the practice of education by helping guide curriculum choices and prospectively assessing representation in the creation of new content. We hope this work will also stimulate a wide range of social science research that uses printed content -- both images and text -- as primary source data to describe patterns of representation and study how variation in this representation shapes human beliefs, behavior, and outcomes. By providing research that expands our understanding about diversity in content, we can help to work to overcome the structural inequality that pervades society and our daily lives.