- Coding:
- Python is not Spotify. Understand all lines.
- But think about if you’re the user or the developer - there’s a tradeoff between understanding the code at a deep level and how fast you want to get it done.
- Read file paths twice before running the code.
- Test it first with one image/tumor/patient before running it for multiple.
- Autosave
- Git
- Celebrate with cheerful messages.
- Experiments:
- What is the task (the question)? What am I learning with this experiment? All experiments should have a take-home message.
- Do I have the data I need?
- How am I going to measure performance? Use more than one metric ideally.
- Computers don’t volunteer information; I have to ask for it.
- Stay optimistic: how can I find something good here?
- Show results to someone before I decide there’s nothing there (you tend to be very negative Olivia!)
- I’m an imaging scientist, visualize the images.
- Better to have bad results than to have no results at all.
- Also better to have output first than to have an optimized pipeline. If performance is an issue after you have some results then optimize pipeline. Goals and outcomes over methods!!
- I can tolerate ambiguity, I’ve done it before. There will always be uncertainty but I can still make decisions based on the best available evidence (healthy skepticism with practical action! pragmatic idealist, not idealist!)
- Take screenshots of failed experiments/plots for the last lab meeting before the viva.
- Its normal that I change my mind after talking to someone I respect but think about it properly adn discuss it with someone else before changing the whole experiement design.
- Celebrate
- Presentations:
- The most important thing is the audience, not me, not the message
- Also: assume people know what cancer is, do not assume they know the rest.
- Productivity
- Productivity is a relationship between input and output. Calibrate expectations.
- Working remotely: respect spaces at home. No laptop on the bed unless I want to deal with the consequences the day after.
- I’m responsible of figuring out how I work best.
- I don’t have to start my day reading my email.
- Too much reading is procrastinating too.
- Document everything & organize information (codes, docs, whatever) by where I will use it, not by where I found it or how I created it.
- Get out of my own head:
- Don’t dwell on the past, stop carrying a cemetery of failed projects/papers in your head.
- I don’t have to say yes. Focus on getting your thesis done. Yes, I have a choice. No, they won’t be upset.
- Everyone is busy, no one is “ignoring” me on purpose, everyone is working on their own goals, and I should be too. The thesis is not everyone’s priority, only mine!
- I’m not my work, I’m a vector for the work.
- I have power over how I react to these thoughts, don’t choose my default setting.
- Meetings:
- They should all have a goal and an agenda.
- Finish with next steps
- Lab meetings are for work in progress, use the opportunity to ask questions and get feedback not to show everything you’ve done.
- Research:
- How can I be useful? (better question than how can I be successful?)
- Being useful for patients and for research are two different things (academic success vs clinical success)
- Read papers from other fields (other primary tumors, imaging modalities, non-imaging AI too..) Someone has thought about it before for sure but perhaps not with imaging data or not for cancer.
- Pen and paper to understand things. Explain it to someone else to test my understanding.
- Ask for help & ask for what I want. Seek out collaborators and mentors. They don’t have to be from my group or even from my institution. There might always be kind, brilliant people out there willing to help me.
- The best resource I have is my mind and the best tool is writing.
- Trátate bien. Making things is hard and my brain wasn’t built to deal with abstract concepts . Also I’m repeatedly and directly confronting the boundaries of your own intelligence and finding it wanting , which is not a pleasant experience. Plus I’m working in English every day. It’s okay to find it hard, calm down.
- It’s a privilege to work with patient data. I have a moral obligation toward them to find if there’s something useful there.
- Random:
- Si se te ocurren ideas cuando vas en moto aparca y apuntátelas, se te van a olvidar cuando llegues a casa.
- No tocarme la cara ni al espalda mientras esté en el ordenador.
- Ideas, data, experiment design.. 99% of it comes from other people I work with. Also people who came before me. Give credit.
- I am a member of different communities (radiomics group, VHIO, LaCaixa) and that comes with responsibilities.
- Don’t start conversations with first year PhDs about mental health and the struggles of the PhD. Exaggerate the good parts as much as the bad parts!!
- Time is not fungible.
- Everyone is trying their best.
- Re-read highlights from the Art of Science and Engineering, Writing your Dissertation in Fifteen Minutes a Day and others.