The Great Instagram Experiment
tldr: In an effort to promote his new book, indie author Anthony Adoré had a few programs developed that automatically parsed his manuscript written in MS Word into images created to post on Instagram.
The human heart is complicated. 001/423. #passion #writersofig #life #prayer #wisdom #qotd #surrender #bookishfeatures #quotes #promote
Between man and woman, there are many complications woven into the fabric of communication. 002/423. #baegoals #jesus #writingcommunity #amreading #happiness #divorce #instabooks #woman #mindful #relationshipadvice
Second, in a practical sense, both speaker and listener, both man and woman, must share the same verbal language. 003/423. #advice #bibleverse #submission #wisdom #introvert #blessed #woman #bibliophile #couples #worship
All authors desire for their works to be read. In order for new authors to announce themselves to the world, they must adopt certain creative strategies to generate exposure. In order to generate maximum exposure with little to no financial investment, I commissioned the development of a Windows 10 program that automatically created “teaser images” of my book manuscript. Over the course of days, weeks, and months, I intend to post these images to my instagram account using popular hashtags.
My “simple” idea was to post the first sentence from every paragraph from my manuscript as an image to social media. In this manner, I would be teasing the book’s content to potential readers, albeit, one sentence at a time. If someone was interested in reading more based on the first sentence from each paragraph, they would, hopefully, purchase the book and discover the connection between all first sentences. Instagram, as we all know, showcases its fair share of quotes as images — and I wanted to follow and blend in with this popular trend. The process of going through my manuscript one paragraph at a time in order to copy-paste the first sentence of each and every paragraph onto an image background using some image processing software would be tedious and time consuming. Fractions of the Soul turned out having around 423 such paragraphs and I wasn’t interested in damaging my fingers with repetitive motions over the keyboard.
I decided to visit one of those sites that allows you to hire a freelance programmer. My initial logic for program design revolved around having a Windows 10 program that would parse a specific MS Word .docx file in order to extract the first sentence of each paragraph and automatically copy-and-paste it onto a 1000px x 1000px background! A few hundred bucks later, I had my program. The program allows me to: (1.) select any MS Word file saved in the .docx format, (2.) choose background image #1 and choose background image #2, (3.) choose the font color, and (4.) choose the font size.
Two directories would be used — “input” and “output.” Inside the input folder, I would have my chosen MS Word document and the two background images. The output folder would store the results. Filenames would be created automatically, e.g., “0001.jpg” and “0001.txt,” and “0002.jpg” and “0002.txt,” etc. The numbering format would follow and correspond to the paragraph number. Therefore file 0001 corresponds and reflects paragraph 1 in the manuscript, and so forth. 0001.txt contains the first sentence from the first paragraph. 0002.txt contains the first sentence of the second paragraph, and so on and so forth until 0423.txt reflects the first sentence of the 423rd paragraph.
The first difficulty I encountered revolved around the use of a period. See, the program indeed scans and finds all first sentences in the specified manuscript and the logic of the program determines that all sentences end with a period. If “e.g.” or “i.e.” or “etc.” was part of the sentence, for example, then the program would end after encountering the first period, making for an incomplete first sentence. To combat this, I had to revise my manuscript and replace all instances of “e.g.,” “i.e.,” and “etc.” with “eg,” “ie,” and “etc” respectively.
The second difficulty I encountered dealt with footnotes. Footnote references provided in the middle of a sentence produced a lot of numerical material not related to the sentence at hand in the final output version. To combat this, I transformed all my footnotes to endnotes and then deleted all the endnotes.
The third difficulty relates to the first and deals with quotation marks. If the first sentence contained quotation marks and also contained a period within those quotation marks, for example, a quote that spans a couple of sentences or more, then the program would output the opening quotation mark but not the closing quotation mark because a period fell somewhere between the two. In the output version, it would seem that a quote started but was not finished.
The fourth and final problem dealt with the length of the first sentence. This isn’t really a problem, but programming logic had to be included to account for potential difficulties. The program would have to detect the length of a sentence in order to determine whether or not to use a smaller font for the entire quote. I selected to use a 14px font size — and when a longer first sentence was detected by the program, the program would use a smaller font size until the entire first sentence fit within the 1000px x 1000px boundaries of the background image.
Randomizing those Hashtags
Apart from using the program, I had to venture off and generate my own hashtags. Using various research methods, I came up with a list of about 160 hashtags that are both hand-picked by me and chosen automatically from a list of popular hashtags. Instagram has a 30 hashtag limit per post. And since I didn’t want to flood the hashtags and appear annoying, I found a site on the internet that allowed me to input my entire list of 160 hashtags and randomly output 10 from the 160. I randomly generated 10 hashtags from the 160 total list 423 times — one generation of 10 tags per image. In this manner, I would not be flooding any hashtag too much.
Remember that the output directory contains 0001.txt. I opened up Notepad++ and on line #1 of Notepad++, I copied and pasted the contents of file 0001.txt AND I copy-pasted the random list of 10 hashtags. I even injected a 001/0423 numbering scheme into the description of each image to ensure the order of images matched with the order of descriptions.
I then uploaded all images to a website that would automatically post each image to my instagram account on my behalf at the same time each day. I chose the specific image to be posted on a specific day followed with the content of the .txt file containing the hashtags and written sentence. The final result is that which you see the Anthony Adoré Instagram Profile.